INTRODUCTION

Innovative therapeutic tools that are more efficient, selective, and economically accessible are essential for the improvement of quality of life (1,2). In the quest for such therapeutic tools, the generation and construction of chemical libraries with high molecular diversities are important (3,4). Because of the rate-limiting character of the hit-to-clinical candidate process, sophisticated catalytic transformations and appropriately planned synthetic strategies should be well established for the delivery of molecules with interesting physicochemical properties (5)(6)(7). Insertion strategies such as single-atom skeletal insertion (8,9), cut-and-sew insertion (10), dearomative insertion (11), and one-or multiatom homologation (12) have been developed as sources to deliver promising candidates for medical applications. Despite substantial progress, the exploration of druglike chemical space to attain synthesizable and innovative structures through the manipulation of polar functional groups is less explored (Fig. 1A) (13). However, such transformations might allow the improvement of key physicochemical characteristics and pharmacokinetic properties (14)(15)(16). For example, the bioactive efficacies were selectively increased upon embedding with additional inert amide groups (17,18). Given the wide applications of polar functional groups in medicinal chemistry (19), the proliferative process to produce more such groups with improved functionalities is an attractive synthetic strategy.

e inert amide, which plays a critical role in the composition of biomolecules (20) and makes up the framework of more than 25% of clinical drugs (21) due to its hydrogen bond donor and acceptor properties (22,23) and inherent high stability (24,25), represents an ideal group for the development of skeletal modification. Recently, ring expansion (26,27), aminoacylation (28)(29)(30)(31), and skeletal elongation (32), as well as genetic code expansion (33,34) by amide C─N bond insertion, have been investigated and used to forge innovative transformations and biofunctionalities. Although widely applied, these tactics entail preactivated or in situ activated amide C─N bonds and require the high efficiency of ribosomal modifications of nonstandard α-amino acids that pose methodological hurdles for diversifications and late-stage applications. us, an efficient and direct amide C─N bond insertion method would be extremely valuable for exploring if interesting building blocks can be easily incorporated into linear and cyclic bespoke frameworks, thereby generating more amides with diverse chemo-and biofunctionalities (Fig. 1B). e anthranilyl unit, an aromatic cyclic β 2,3 -amino acid residue, is a privileged scaffold widely existed in bioactive molecules (Fig. 1C) (35)(36)(37)(38). Direct insertion of such unit through the mentioned conversion would open an economical and proliferative process to achieve reshaped oligoamides and strained amide-containing macrocyclic architectures (39,40).

Anthranilium salts, which are always generated from the alkylation of their parent molecules with oxonium or carboxonium, have been used in versatile syntheses to construct anthranilyl frameworks by the in situ generation of reactive iminoketene intermediates (41)(42)(43). However, the direct application of such intermediates inserted into valuable structural motifs remains challenging and necessitates effective strategies. Considering the potential acyl and amine donor and acceptor abilities of this iminoketene intermediate, if amide was easily trapped, we envisioned that hydrogen bonding interactions would probably be formed because of amide and its enol tautomer unique hydrogen bond donor and acceptor characteristics. Subsequent acyl moiety fixation through this bonding would induce amide C─N bond deconstruction and enable anthranilyl unit insertion to achieve amide proliferation. Herein, we report an iminoketeneinduced proliferation process through the insertion of anthranilyl scaffold into versatile amides to furnish valuable oligoamides and cyclic peptides (Fig. 1D). Given the synthetic challenges of amides (44)(45)(46), especially selective amide N-alkylation (47), this process not only facilitates the repurpose of amide by circumventing the laborious de novo synthesis but also enables versatile amide-containing pharmaceuticals, natural products, and bioactive molecules as points of proliferation for generating innovative analogs through the latestage editing modality.

of 10

RESULTS

Reaction development and mechanistic studies

To verify the feasibility of the anthranilyl unit incorporation process, we commenced this study by using the inert amide S1 and Nmethylanthranilium fluoroborate (NMAFB, S2) as the substrates. A er extensive screening of reaction conditions (Fig. 2A), we identified the combination of triethylamine (Et 3 N) as the base and dichloromethane as the solvent at 60°C as the optimal reaction conditions (see table S1 for more condition optimization details), which afforded the oligoamide 1 with a very good yield (80%). Owing to steric hindrance, the two adjacent amide subunits always rotate out of planarity, affording hydrogen bond formation (48). If N-alkylation selectively takes place, the N,N-substituted amide features the low-energy rotameric cis/trans barrier due to hydrogen bond dissociation (49, 50), which could allow for the variation of cis/trans ratios in different deuterated solvents (for details, see fig. S1). Control experiments indicated that 3-substituted NMAFB cannot enable the formation of insertion product 1a, whereas the preactivated amides show similar reaction behavior under standard conditions (Fig. 2B). ese results clearly support that the formation of the iminoketene intermediate and the NH group of inert amide plays vital roles in C─N bond deconstruction and reconstruction. To explore the nature of the anthranilyl unit insertion process, density functional theory calculations (see the Supplementary Materials for computational details) were performed on the proposed mechanism of anthranilium salts-mediated amide C─N bond recombination (Fig. 2C). e ring-opening of anthranilium salt to afford iminoketene is exergonic by -26.1 kcal/mol. e nucleophilic addition of S1 to this active species via transition state (TS-1, Downloaded from https://www.science.org at University of California Los Angeles on August 20, 2025

∆G ‡ = 15.0 kcal/mol) gives anthranilamide. With the presence of another molecule S1, which participates in intermolecular hydrogen bond formation, a ring-closure intermediate INT3 is formed through the transition state (TS-3, ∆G ‡ = 17.6 kcal/mol). erea er, the C─N bond cleavages in accordance with hydrogen transfer through the transition state (TS-4, ∆G ‡ = 27.6 kcal/mol), which is generated from hydrogen bonding isomerization of INT3, giving the thermodynamically stable product 1. In the absence of another S1, proliferation from INT1 has a higher activation energy barrier (TS-2, ∆G ‡ = 29.8 kcal/ mol) due to the lack of stabilization from intermolecular hydrogen bonds. ese results show that the hydrogen bond network is crucial to the C─N bond proliferation and amide acts as starting reagent and catalyst in this insertion process.

Substrate scope exploration

With the optimized conditions in hand, the scope of the inert amides and anthranilium salts was explored (Fig. 3A). Our study primarily examined the broad applicability of linear inert amides, resulting in the construction of a range of oligoamides with moderate to excellent yields. ese products encompassed a diverse array of motifs, including both naturally occurring and pharmaceutically notable building blocks. For example, diverse aromatic amides (2-24) and aliphatic amides (25)(26)(27)(28)(29)(30)(31)(32)(33) were all well compatible with this insertion procedure. e transformation proceeded smoothly using the salts bearing different substituents or extended aromatic system to form the corresponding products in high yields (34)(35)(36)(37)(38)(39)(40)(41)(42)(43)(44). Beyond aromatic and aliphatic amides, other types of inert amides (such as inert thioamide, N-benzoylbenzamide, and secondary αketoamides) could be transformed into the desired oligoamides through the established C─N bond deconstruction and reconstruction process (45)(46)(47)(48)(49)(50)(51)(52)(53). Substrates featuring biologically relevant scaffolds, which were derived from monosaccharides, disaccharide, and amino acids, were all successfully proliferated with high efficiencies (54)(55)(56)(57)(58)(59)(60)(61)(62). From the perspective of product diversity, primary amides were also explored, leading to functionalized quinazolinones (63-67) through consecutive scaffold insertion and intramolecular dehydration process (Fig. 3B; see table S3 for condition optimization details). To further illustrate the utility of this transformation, subsequent work was demonstrated by using complex molecules as starting materials in medicinally relevant settings. We found that complex frameworks, which feature distinct physicochemical properties and are generated via molecular hybridization from the combination of natural product with drug, drug with drug, natural product with natural product, and drugs with pharmacophore, were all suitable for the explored insertion process and quickly achieved increased molecular complexity (68-71) (Fig. 3C).

Late-stage amide proliferation

Encouraged by the above results, we subsequently performed the skeletal editing to provide a general method for installing anthranilyl scaffold directly into various kinds of bioactive compounds, ranging from pharmaceuticals, metabolite, and natural products to a panel of agrochemicals and veterinary drug in Fig. 4 (72-100), thereby avoiding tedious and costly de novo synthesis. Versatile ubiquitous N-,O-, S-, and C-containing (hetero)cycles and sensitive functional groups in medicinally relevant fields were well tolerated and afforded the corresponding products in acceptable-to-good yields. Drugs, natural products, and veterinary medicine with complex structures, such as paclitaxel (a er esterification), bimatoprost (a er esterification), lappaconitine, and eprinomectin (a er esterification), were all engaged in this proliferation to assemble the corresponding derivatives in synthetically useful to moderate yields without erosive of the rest molecule framework (84, 88, 96, and 100). e dual insertion route in one pot to assemble cabozantinibbased oligoamide was also successful and delivered the proliferated product with high efficiency (86). Nirmatrelvir (the major component found in paxlovid) and chlorantraniliprole (agrochemical) were specifically converted to corresponding products (92, 99), in which the reaction occurred in non-crowded structural environments while leaving sterically hindered amides and other functional groups untouched. Aliphatic inert amide-containing drugs, such as agomelatine (an anti-insomnia drug), nateglinide (a noninsulin-dependent antidiabetic drug a er esterification), and linezolid (an antibiotic used to treat Gram-positive bacterial infections), were also tolerated for this skeletal insertion process and afforded the recombination products without affecting the stereoselectivities (75, 78, and 79). Indibulin, an α-ketoamide-connected anticancer drug, was additionally compatible with the developed protocol, and the corresponding derivative was obtained in good yield (85). When roflumilast was subjected to the established synthetic system at gram scale, the transformation proceeded smoothly with slightly affecting the efficiency (81, 4 mmol, 1.42 g, 66%). On the basis of the rapid construction of chemical libraries with high molecular diversities through the established insertion strategy, analogs derived from amide-containing bioactive molecules under mild conditions are amenable to high-throughput screening, systematic structure-activity relationship exploration, and fragment-based drug discovery campaigns.

Cyclic peptides syntheses

Considering the widespread occurrence of cyclic amides in bioactive molecules and pharmaceuticals, the skeletal editing strategy for such amides with different ring sizes has also been explored with anthranilium salts at room temperature (Fig. 5A, and see table S2 for condition optimization details). e cyclic substrates bearing different functional groups and scaffolds were smoothly proceeded and afforded the 8 to 25-membered cyclic peptides through this insertion strategy (101-113, 34-81% yield, cis/trans 1:1 ≥ 20:1). e eight-membered peptide 102 was assembled through a ring expansion and desulfonylation process using 4-(phenylsulfonyl)-2-azetidi none as the starting substrate under standard conditions. To further highlight the robustness of our strategy, the skeletal editing of cyclic amide-containing drugs with anthranilium salts under standard conditions was also explored. As expected, cyclic peptides generated from phenidone B, pirenzepine, nevirapine, and repotrectinib were all achieved with high efficiencies (114-117) (Fig. 5B). Notably, a gram-scale experiment implicated the reliability and practical applicability of this approach, yielding the desired product 103 with 55% yield.

Synthetic transformations and applications

On the basis of the impressive biological profile generated from the anthranilic diamide scaffold and its analogs, to further assess the value of this protocol, site-specific N-methylation and Ndemethylation were also evaluated. As showcased in Fig. 6A, Nmethylated tariquidar (118), a P-glycoprotein drug e ux pump inhibitor of its original molecule, was rapidly synthesized using the established editing strategy. Conversely, N-demethylated anthranilic diamide 119 was efficiently assembled from roflumilast-based oligoamide 81, which could potentially serve as an apoptosis inducer to achieve anticancer activity (51,52). e utility of the established insertion procedure was further extended by the efficient synthesis of natural products. Terremide A (123) and terremide D (124), the secondary metabolites from marine-derived fungi (53,54), were concisely first synthesized within three or two steps from easily accessible starting materials (Fig. 6B). To further showcase the synthetic potential of this insertion, we conducted the rapid syntheses of several bioactive molecules, with the objective of directly decreasing the overall step count. As illustrated in Fig. 6C, the BCL-3 inhibitor, which was previously assembled in two steps in less than 10% overall yield (55), could be more efficiently accessed by directly inserting the anthranilyl scaffold into related readily available starting amide (125, 45%). e agrochemical 126 has certain insecticidal activities against aphids, cotton bollworms, armyworms, and beet armyworms, which required a four-step synthesis with a 14% overall yield (56). Using the insertion strategy, this insecticide could be more efficiently constructed in a single step with a 54% yield, representing a dramatic improvement over the established synthesis. Furthermore, this editing strategy provided an ideal pathway for the direct elongation of linear peptides by the insertion of the anthranilyl unit into inert amides (Fig. 6D). Starting with a dipeptide, a linear artificial tripeptide 127 was rapidly prepared with a 28% yield. e value of this editing protocol was also extended through the synthesis of oligoanthranilamide 128 by using 1 as a starting material. Considering cyclic amide reshaped within this insertion process, we Downloaded from https://www.science.org at University of California Los Angeles on August 20, 2025

wondered whether our editing protocol could also enable successive ring expansion through multi-anthranilyl scaffold insertion. We found that such methods have the power to expedite the assembly of functionalized peptide macrocycles through the selective, iterative insertion of anthranilyl scaffolds into various cyclic amide-containing molecules. For example, 13-membered lactam and 18-membered binaphthyl-containing diamide could undergo several iterations through the same chemistry to construct 17-, 21-, 22-, 25-, and 26-membered macrocyclic peptides (129-133), respectively. is ring expansion process could be scaled (10 mmol) with slight erosion of the yield (129, 34%). In addition, a 25-membered antibiotic (rifamycin S a er esterification) was compatible with this sequential protocol, enabling the synthesis of 29-and 33-membered peptide macrocycles 134 and 135 via successive skeletal expansion. Considering the versatility of this insertion method, it is conceivable that versatile libraries of derivatives can be readily constructed for further medicinal investigations.

DISCUSSION

We have developed a viable synthetic strategy for the rapid preparation of oligoamides and cyclic peptides through amide C─N bond proliferation. e explored methodology is exceptionally mild and tolerates versatile functional groups and diverse biologically important scaffolds. is characteristic enables its compatibility with the skeletal editing of a wide range of drugs, natural products, and agrochemicals via anthranilyl scaffold insertion. is insertion method could also serve as a pivotal synthetic strategy for constructing various cyclic peptides for further diverse chemo-and bio-applications. In addition, the merits of this protocol are extended by, to our knowledge, the first syntheses of marine natural products, the rapid syntheses of bioactive molecules, and the efficient assemblies of functionalized peptide macrocycles. Given the prevalence of amides in biologically relevant molecules, we believe that this incorporation strategy will find widespread applications in contemporary drug discovery and development.

MATERIALS AND METHODS

General procedure for the proliferation of secondary amides

To a dried Schlenk vial were added secondary amides (0.3 mmol, 1.5 equiv) and N-alkylated anthranilium fluoroborates (0.2 mmol, 1.0 equiv), followed by the addition of dry CH 2 Cl 2 (1 ml) and Et 3 N (56 μl, 0.4 mmol, 2.0 equiv) in a glove box. e reaction was refluxed at 60°C under argon atmosphere overnight. A er being purified by flash column chromatography on silica gel, the desired product was obtained.

General procedure for cyclic peptides synthesis through the proliferation strategy

To a dried Schlenk vial were added cyclic amides (0.2 mmol, 1.0 equiv) and N-alkylated anthranilium fluoroborate (0.3 mmol, 1.5 equiv), followed by the addition of dry PhOMe (1 ml) and TMEDA (90 μl, 0.6 mmol, 3.0 equiv) in a glove box. e reaction was stirred at room temperature under argon atmosphere overnight. A er purification by flash column chromatography on silica gel, the desired product was obtained.