Merging photocatalysis with C-H insertions enabled switchable synthesis of indoles and indolines

Abstract

C–H insertion offers a powerful route to C–C bond formation, yet it has often proven challenging to integrate into cascade synthetic sequences for further molecular elaboration. This limitation stems primarily from the aggressive conditions required for carbene generation and the high sensitivity of carbene species, which lead to premature termination and incapability with downstream transformations. Although intramolecular C–H insertion provides direct access to indolines, conventional approaches tend to produce ungoverned mixtures of indolines and indoles, owing to harsh conditions. To address these limitations, we develop a switchable synthetic platform that enables tunable access to either scaffold. By leveraging energy transfer, C-H insertions, and consecutive photoredox dehydrogenation, triplet carbene is accommodated in cascade reactions to afford N-fused indoles selectively. Conversely, the strategy can be switched to access indolines via singlet carbene-mediated C-H insertion. The incorporation of both motifs into drug scaffolds and natural products demonstrates their practical applicability. Based on the construction of a pseudo natural product (PNP) library featuring novel indole and indoline derivatives, tetracyclic indole alkaloid 5b is identified as a promising ferroptosis inhibitor after comprehensive pharmacological studies, underscoring the potential of PNP synthesis in drug discovery.

Intramolecular C– H insertion provides direct access to indolines, but conventional approaches tend to produce ungoverned mixtures of indolines and indoles, owing to harsh conditions. Herein, the authors report a switchable synthetic platform that enables tunable access to either scaffold.

Introduction

Carbenoid C-H insertions have garnered considerable interest in organic synthesis^1,2 and medicinal chemistry^3,4, owing to their potential for direct C–C bond formation. Physical progress has been made in accessing carbenoid C-H properties^5–12. Among these, transition-metal-catalyzed decomposition of stabilized diazo compounds has emerged as a highly effective method, enabling insertion into a variety of C–H bonds, including alkyl and aryl substrates^13–20. The scope of carbene chemistry has been further expanded through the use of alternative diazo precursors such as N-tosylhydrazones, which undergo efficient C–H insertion via donor/donor carbene intermediates to afford diverse C–C coupled products (Fig. 1a)^21–33. In addition, the rapid development of photochemical synthesis has opened new avenues for carbene generation under mild conditions. Photoinduced or photocatalytic activation of diazo compounds or N-tosylhydrazones has enabled not only classical carbenoid insertions but also previously inaccessible transformations, as pioneered by the group of König^34,35, Davies³⁶, Koenig^37,38, Sun³⁹ and others^40–43. Notably, the photoinduced C-H insertions have been achieved by our group recently using non-activated alkanes/ethers as coupling partners⁴⁴. The Davies group has also found the C-H insertions via direct photolysis of aryldiazoacetates³⁶. Despite these achievements, the resulting products from C-H insertion are often difficult to further synthetic transformation under cascade settings. This limitation stems primarily from the aggressive reaction conditions required for carbene generation and the high sensitivity of carbene species, which always cause premature termination and the introduction of functionalities incompatible with downstream transformations.

Fig. 1. Controllable synthesis of indoles and indolines.

a Representative C-H insertions and challenges. b State of the art for delivering indoline/indole mixtures. c This work: switchable synthesis of indolines and indoles using a photocatalytic platform.

N-Fused indolines and indoles represent prevalent structural motifs frequently found in natural products^45–48. It was also known that the construction of pseudo natural products (PNPs) has been the emerging method to access the chemical evolution of natural products^49–51. Therefore, the development of synthetic methodologies becomes significant to achieve the diversification of PNPs, including N-fused indolines and indoles. Traditionally, the synthesis of indolines from intramolecular C–H insertion has been advanced by the group of Kehler, which required the microwave irradiation (135 °C)⁵². Later, the Fillion group reported the indoline synthesis, which relied on transition-metal catalysis and high reaction temperatures (100 °C). Notably, the Fillion group observed minor indole side products in some cases⁵³. However, the reactivity was inefficient and yielded an uncontrolled mixture of indoles and indolines. Building on this prior research, the Takai group demonstrated that Mo(CO)₆ and ortho-quinone could proceed the direct deoxygenative insertion of carbonyl carbon. However, this method required harsh heating conditions (135–160 °C) and lead to uncontrolled partial dehydrogenation, resulting in an unselective mixture of indoline (68%) and indole (22%) (Fig. 1b). A subsequent stoichiometric oxidation step was therefore necessary to achieve full aromatization⁵⁴. Although recent advances have been made in the switchable synthesis of related N-heterocycles via intermolecular assembly strategies^55,56, applicable strategies enabling the switchable synthesis of indoles and indolines via intramolecular C-H insertions retained large unexplored. Thus, the development of a switchable synthetic platform combining C–H insertion with complementary strategies would constitute a considerable advance in synthetic methodology for achieving programmable access to either scaffold and facilitate the construction of PNPs and related libraries.

Ferroptosis holds significant therapeutic potential for addressing therapy-resistant cancers, ischemic organ damage, and other degenerative diseases driven by lipid peroxidation⁵⁷. Current small-molecule inhibitors of ferroptosis, including aromatic amines, phenols, ubiquinol, and others, have demonstrated considerable potential in research and therapeutics^58–61. Nevertheless, existing inhibitors still suffer from drawbacks such as poor stability, low specificity, insufficient efficacy, and difficult druggability. Consequently, the development of potent and selective ferroptosis-modulating lead compounds represents a critical unmet need in drug discovery⁶². In this work, building upon these findings and driven by our ongoing efforts in photochemical activation of N-tosylhydrazones as carbene precursors^63–66 and involvement in ferroptosis research^67–69, we envision a strategy for the switchable synthesis of indolines and indoles through the merger of photoredox catalysis and carbene-mediated C–H insertion. At first, we postulate that the triplet carbene from in-situ formed diazo compounds is formed under energy transfer, which further undergoes the intramolecular C-H insertion to afford the indolines. Subsequently, the photoredox cycle may be feasible to facilitate the dehydrogenation of formed indolines, thereby obtaining the desirable indoles without the need of exogenous oxidants and extra reaction steps. Moreover, removal of the photocatalyst furnishes indolines exclusively, avoiding the formation of ungoverned product mixtures (Fig. 1c). We further apply this switchable protocol to synthesize drug-like and natural product analogs. Finally, compound 5b is identified as a promising ferroptosis inhibitor, underscoring its potential for therapeutic development.

Results

Reaction optimization

Our investigation commenced by preparing a N-tosyhydrazone 1a as starting material. After the systematic optimization of different photocatalysts, we found that using 3DPAFIPN as a photocatalyst, sodium tert-butoxide as base, and DCE as the solvent afford the indole 1b in 92% isolated yield under irradiation with 456 nm light for 12 h (Fig. 2). Remarkably, only 3% indoline was observed in this system, demonstrating the high selectivity of this system. Different solvents were further screened and 1,2-dichloroethane (DCE) as the optimal reaction medium. Subsequent evaluation of bases revealed sodium tert-butoxide (NaOtBu) as the most effective promoter, achieving excellent conversion efficiency under standard conditions. Finally, control experiments confirmed the indispensable roles of the catalyst, light, and base in enabling the reaction system for synthesizing indole derivatives. Subsequently, to switch the pathway towards indolines 1c, we sought conditions that arrest the cascade at the insertion stage. It was determined that using 1,4-dioxane as a solvent provided superior reaction efficiency (87%) compared to other solvents (acetonitrile, 1,2-dichloroethane, tetrahydrofuran, fluorobenzene). DBU proved to be the optimal base, and evaluation of light sources identified 427 nm monochromatic irradiation as optimal. Control experiments further confirmed the absolute necessity of both light irradiation and the presence of base. Notably, the optimized indoline-forming conditions demonstrated complete chemoselectivity, with no detectable formation of indole byproducts.

Fig. 2. Optimization and control experiments.

a General reaction conditions A: a (0.1 mmol), base (3.0 equiv.), solvent (0.1 M, 1.0 mL), PC (2 mol%), Kessil lamps (456 nm, 40 W), room temperature, 12 h. General reaction conditions B: a (0.1 mmol), base (1.5 equiv.), solvent (0.1 M, 1.0 mL), 40 W 427 nm Kessil lamps, room temperature, 12 h, stirred at 300 rpm. b Yields were determined by ¹H NMR analysis using 1,3,5-trimethoxybenzene as an internal standard. n.d. not detected. 3DPAFIPN 2,4,6-tris(diphenylamino)−5-fluoroisophthalonitrile, DBU 1,8-diazabicyclo[5.4.0]undec-7-ene, DBN 1,5-diazabicyclo[4.3.0]non-5-ene, DCE 1,2-dichloroethane, NaOtBu sodium tert-butoxide.

Substrate scope analysis

To demonstrate the systematic nature and broad applicability of this switchable strategy, we investigated the substrate scope for the selective synthesis of indoles under the optimized conditions in Fig. 3. At first, N-tosylhydrazones derived from different amines demonstrated good tolerance in this strategy. Six-membered cyclic amines, including piperidine and its mono-substituted derivatives, as well as 1,2,3,4-tetrahydroisoquinoline and its mono-substituted analogs, all of which exhibited excellent compatibility with the methodology to afford the target products in moderate to excellent yields (1b-7b). The reaction process exhibited remarkable insensitivity to the electronic nature or steric bulk of substituents on amine derivatives, demonstrating broad functional group compatibility. This substituent-independent behavior allowed for efficient C-H functionalization across diverse amine substrates without significantly compromising reaction efficiency or product yield. Moreover, the protocol demonstrated remarkable adaptability to ring size variation: whether contracting to five-membered rings or expanding to seven-, eight-, and nine-membered cyclic systems, the reactions consistently delivered the desired N-fused polycyclic indoles with high efficiency (8b-14b). Notably, linear aliphatic amines also demonstrated excellent adaptability in this protocol, achieving the target product with over 90% yield (15b, 16b). Moreover, this redox platform also accommodated N-tosylhydrazones generated from aryl-substituted acetophenones (varying in electronic and steric properties), delivering structurally diverse products in 76-89% isolated yields (17b-20b). The N-tosylhydrazone derived from phenylacetaldehyde afforded the target product 21b in 53% yield. Moreover, N-tosylhydrazones bearing other substituents (ethyl, cyclopentyl, phenyl) were demonstrated to be viable substrates, delivering corresponding polycyclic indoles in moderate to good yields (22b-24b).

Fig. 3. Substrate scope for constructing N-fused polycyclic indoles.

General reaction conditions A: a (0.1 mmol), NaOtBu (3.0 equiv.), DCE (0.1 M, 1.0 mL), 3DPAFIPN (2 mol%), Kessil lamps (456 nm, 40 W), room temperature, 12 h, stirred at 300 rpm. Isolated yield. dr diastereomeric ratio, equiv. equivalent.

Underscoring the versatility of this switchable protocol, we subsequently explored the divergent synthesis of indolines (Fig. 4). Initial exploration of various amines revealed that both piperidine and para-monosubstituted piperidine derivatives afforded the corresponding C-H insertion products in good yields (1c-3c). Notably, spirocyclic and bridged piperidine derivatives were well tolerated under the reaction conditions (4c-6c). The protocol also proved effective for piperazine derivatives, delivering desired polycyclic indolines with satisfactory yields (7c, 8c). Furthermore, the synthesis of the corresponding products was successfully achieved using thiomorpholine, morpholine, and their substituted analogs, all of which demonstrated moderate to good reactivity (9c-11c). Regrettably, five-membered cyclic and acyclic amines proved incompatible with the current system (Fig. S2). To further evaluate the generality of this strategy, we examined the scope of aryl N-tosylhydrazones. Both mono- and di-substituted aryl N-tosylhydrazones participated effectively in the transformation, providing intramolecular C-H insertion products with excellent yields (12c-15c). The benzaldehyde-derived N-tosylhydrazone reacted smoothly to afford the corresponding product 16c. The system exhibited remarkable compatibility with N-tosylhydrazones bearing diverse substituents, including phenyl and ethyl groups, all of which generated corresponding products efficiently (17c, 18c). Collectively, this facile and efficient protocol has significantly extended the scope of substrates for indoline synthesis through photoinduced intramolecular C-H insertion.

Fig. 4. Substrate generality for the synthesis of N-fused indolines.

General reaction conditions B: a (0.1 mmol), DBU (1.5 equiv.), 1,4-diaxane (0.1 M, 1.0 mL), 40 W 427 nm Kessil lamps, room temperature, 12 h, stirred at 300 rpm. Isolated yield. dr diastereomeric ratio, equiv. equivalent.

Synthetic applications

To comprehensively evaluate the practical utility of this controllable strategy, we extended its application to encompass structurally diverse pharmaceutical agents, focusing on late-stage functionalization of drug molecules and natural products (Fig. 5). This dual-purpose methodology was successfully employed for the targeted modification of desloratadine (25b, 19c) (a second-generation antihistamine), enabling precise site-selective transformations. This strategy was successfully applied to the synthesis of compound 26b, a key precursor that can be converted into the natural product cryptaustoline via a simple transformation^70,71. Additionally, the core scaffold of the atypical antipsychotic risperidone (27b) was also constructed. Furthermore, the intramolecular C-H insertion protocol demonstrated robust performance in the structural elaboration of cyclizine (an antiemetic agent), perospirone (a serotonin-dopamine antagonist), vortioxetine (an antidepressant antagonist), paroxedtine (an antidepressant antagonist), providing corresponding indoline derivatives (20c-24c) with excellent efficiency. These collective achievements underscore the transformative potential of this strategy in the construction of PNPs and natural products research. To further demonstrate the practical applicability of this methodology, gram-scale syntheses were successfully executed. As summarized in Fig. 5, using substrate 1a at a 7.0 mmol scale under two distinct sets of reaction conditions afforded the corresponding products 1b and 1c in yields of 77% (1.02 g) and 69% (0.88 g), respectively. Notably, the obtained products were both commercially unavailable, highlighting the synthetic potential of this method. Moreover, gram-scale synthesis of commercially prohibitive compound 14b was achieved in 73% yield (1.6 g, US$69/mg), demonstrating this platform’s capacity for cost-effective and scalable production. Accordingly, a PNP library centered around novel indole and indoline scaffolds was established.

Fig. 5. Late-stage modification of drug molecules/natural products and gram-scale synthesis.

dr diastereomeric ratio, equiv. equivalent.

Mechanistic studies

Finally, a series of mechanistic experiments was performed to elucidate the mechanism underlying this strategy. The time course revealed that 12 hours were required to achieve satisfactory yields of indoles and indolines (Fig. 6a). On-off experiments demonstrated that light is crucial for proceeding the synthesis of both products (Fig. 6b). UV-Vis spectroscopy revealed no significant red- or blue-shift in the hydrazone absorption peak upon NaOtBu addition, demonstrating the absence of complex formation (Fig. S5). However, the visible absorption spectrum revealed that the additive DBU induced an obvious bathochromic shift, suggesting the formation of a non-covalent complex between the DBU and N-tosylhydrazone anion (Fig. 6c)⁶⁴. Stern-Volmer analysis indicated that the excited state of the photocatalyst was quenched by the N-tosylhydrazone anion (Fig. 6d). Cyclic voltammetry further suggested that the excited state of 4CzIPN (E_red = −1.61 V vs Ag/AgCl) possesses sufficient reduction potential to reduce DCE (E_red = −1.5 V vs Ag/AgCl), which was also consistent with a previous report (Fig. 6e)⁷². While these results support a SET mechanism, an amine-mediated halogen-atom transfer (XAT) pathway remains a plausible alternative. It is noteworthy that indoline underwent transformation to indole under standard catalytic conditions (37% yield), indicating it served as a reaction intermediate in the indole formation sequence (Fig. 6f). Moreover, radical trapping experiments with TEMPO and BHT caused drastic yield suppression in indole formation (Fig. 6g), implicating radical involvement in the synthesis of indoles. Furthermore, perylene addition as a triplet carbene quenchant strongly inhibited reactivity, indicating that indole synthesis likely proceeds through an energy transfer pathway to generate triplet carbene intermediates. Conversely, indoline formation via C-H insertion showed no significant yield decrease with radical quenchers, excluding the radical process. At last, carbene trapping with styrene and phenylacetylene afforded cyclopropanation products, as confirmed by HRMS, which is consistent with prior reports (Fig. 6h)^73,74.

Fig. 6. Mechanistic studies.

a Time course experiment. b On-off experiment. c Uv-vis absorption spectrum. d Stern-Volmer plot: photocatalyst was quenched by the N-tosylhydrazone anion 1a’. e Cyclic voltammetry. f N-Fused indole was afforded from indoline under standard conditions. g Radical quenching experiments and triplet quenching experiments. h Carbene trapping experiments.

To further understand the reaction mechanism of this photocatalytic C–H insertion, DFT calculations were performed at the UM06-2X/def2-TZVP-SMD(DCE)//UM06-2X/def2-SVP-SMD(DCE) level of theory (Fig. 7a). The singlet N-tosylhydrazone anion Int1a(S), which serves as the primary quencher of the photocatalyst (Fig. 6d), was employed as the starting point for the free energy surface calculations. Upon energy transfer from the photocatalyst, Int1a is excited to the triplet state, resulting in an energy increase of 45.3 kcal/mol. The triplet N-tosylhydrazone anion Int1a(T) then undergoes sulfonate loss via transition state TS2a(T) to form the triplet diazo species Int3a(T), overcoming a free energy barrier of merely 7.8 kcal/mol. The diazo species Int3a(T) subsequently releases nitrogen through transition state TS4a(T), leading to the formation of the triplet carbene species Int5a(T), which can be substantiated by quenching with a radical scavenger (Fig. 6g). Int5a(T) rapidly undergoes conformational change via TS6(T), with a free energy barrier of 3.9 kcal/mol, to generate the Int5b(T) conformation that is prone to intramolecular reaction. The significance of such conformational changes in influencing reaction pathways has also been demonstrated in our previous work⁷⁵. Int5b(T) then undergoes hydrogen atom transfer (HAT) through transition state TS7(T), generating the intramolecular triplet diradical species Int8(T). Subsequently, Int8(T) can undergo an intersystem crossing (ISC) to form the intramolecular zwitterionic species Int8(S). This ISC via MECP2 requires an electronic barrier of 4.3 kcal/mol through the located minimum energy crossing point (MECP). Further intramolecular cyclization of the zwitterionic Int8(S) results in the formation of the product Int10(S). We also calculated the direct sulfonyl release from the singlet N-tosylhydrazone anion Int1a(S), which requires overcoming a free energy barrier of 32.5 kcal/mol via transition state TS2a(S). This indicates that the singlet-state pathway is kinetically inaccessible under standard conditions, thus highlighting the indispensable role of photocatalysis in facilitating this transformation. Additionally, we explored the possibility of the reaction proceeding through the singlet carbene species Int5b(S), which can be generated from the triplet carbene species Int5b(T) via ISC through MECP1, in accordance with the carbene capture mechanistic experiments (Fig. 6h). The transition state TS7(S) is only 0.6 kcal/mol higher in energy than the hydrogen atom transfer transition state TS7(T), indicating that C–H insertion through singlet carbene is also an operative pathway in the reaction. These calculations further substantiate our understanding of the reaction mechanism of photocatalytic C–H insertion proceeding through the triplet state.

Fig. 7. The capture of intermediate and reaction energy profiles.

a DFT-computed free energy profile for the photocatalytic C-H insertion of N-tosylhydrazone anion. b Plausible mechanism.

Based on the above evidence, a plausible mechanism was proposed (Fig. 7b). The indoline was first generated through the triplet carbene-mediated C-H insertions, which featured that the triplet carbene species could undergo the HAT process to form the diradical species. The formed indoline 1c was further oxidized by the excited state of the photocatalyst to generate the radical cation 1 g, which was subsequently deprotonated to form the radical species 1 h. Meanwhile, the DCE as an electron donor, was reduced by the radical anion of PC. Afterwards, the 1 h was further oxidized to generate the cation 1i, which was deprotonated again to form the desirable indole 1b (Path A). In addition, the indoline was also selectively afforded under the photoinduced conditions without the additional photocatalyst. The activation of N-tosylhydrazone was achieved, and the donor/donor diazo intermediate was formed through the formation of non-covalent interactions. The diazo intermediate further released the nitrogen via photolysis to generate the singlet carbene species Int5b(S), which further undergoes the C-H insertion to afford the indoline (Path B). It should be noted that no indole was detected under these conditions due to the relatively mild conditions without thermal conditions or strong base, presenting the advances of our protocols compared with the previous reports.

Biological evaluation

To identify candidate drugs with anti-ferroptosis potential, we first established a classical model of ferroptotic cell death induced by RSL3, which directly inhibits glutathione peroxidase 4 (GPX4), in ferroptosis-sensitive HT-1080 human fibrosarcoma cells. Based on this, we used a cell viability assay to screen our PNPs library with indole and indoline compounds at a final concentration of 10 μM, and selected three compounds, namely 5b, 11c, and 16c, which displayed the strongest protection against RSL3-induced cell death (Fig. 8a). We then validated the anti-ferroptosis activity of the three candidate drugs in a dose-dependent manner and compared them with the ferroptosis inhibitor Ferrostatin-1 (Fer-1). As shown in Fig. 8b, 5b at 25 μM almost fully rescued RSL3-induced cell death, while 16c or 11c at the same concentration did not show such significant protection against RSL3. The toxicity of the three candidate drugs is similar, with IC₅₀ values between 300-320 μM, which means that there is basically no toxicity at the working concentration (Fig. 8c). Imidazole ketone erastin (IKE), an arbonyl erastin analog, can induce ferroptosis through inhibiting the cystine transporter SLC7A11, resulting in cysteine deprivation, subsequent glutathione (GSH) depletion, and ultimately GPX4 inactivation. Similar to Fer-1, 5b treatment also significantly protected HT-1080 cells against IKE-induced ferroptosis, suggesting that it has the ability to confer protection against ferroptosis elicited by glutathione depletion (Fig. S10). Lipid peroxidation is the core mechanism involved in ferroptosis. Consistently, we found that 5b could markedly block lipid peroxidation induced by RSL3 or IKE in HT1080 cells (Fig. 8d). Taken together, these findings indicate that 5b is a potential ferroptosis inhibitor. It is well-known that tetracyclic indole alkaloids hold a pivotal and leading role in natural product-based drug discovery. Therefore, these findings demonstrate that constructing a focused library of PNPs based on the tetracyclic indole alkaloid scaffold could be a highly productive strategy for drug discovery.

Fig. 8. Research on the potential anti-ferroptosis capability.

a Screening of indole and indoline using a cell viability assay. b Anti-ferroptosis activity of three candidate compounds (n = 4 per group). c Cytotoxicity assessment of three candidate compounds (n = 4 per group). d Lipid peroxidation levels were tested by using flow cytometry in cells treated with RSL3 or IKE, combined with three candidate compounds (n = 3 per group). For (b–c), n independent experimental samples; data were presented as mean ± SEM. Statistical significance was determined by a two-tailed Student t-test.

Discussion

In summary, we have developed a photocatalytic platform that synthesizes indoles by harnessing the reactivity of triplet carbenes in cascade reactions through the integration of energy transfer, photoredox catalysis, and C–H insertion. Alternatively, this switchable protocol affords efficient access to indolines without subsequent dehydrogenation. The practical utility of this methodology was demonstrated through late-stage modification of pharmaceutical scaffolds and natural product frameworks bearing indoline/indole motifs, as well as through successful gram-scale implementation. Finally, combined mechanistic investigations and density functional theory (DFT) calculations corroborate the involvement of energy transfer, C–H insertion, and photoredox dehydrogenation in this process. Furthermore, tetracyclic indole alkaloid 5b, identified as a promising ferroptosis inhibitor, is currently under extensive investigation in our laboratory.

Methods

General procedure for the synthesis of N-fused indole

A dry 10 mL Schlenk tube containing a stirring bar was charged with N-tosylhydrazone (0.1 mmol, 1.0 equiv.), 3DPAFIPN (2 mol%), NaOtBu (0.3 mmol, 3.0 equiv.). After purging the flask three times under vacuum and three times with argon, DCE (1.0 mL, 0.1 M) was added while maintaining an argon atmosphere. The reaction was carried out under Kessil lamps (456 nm, 40 W) for 12 hours in a reaction apparatus (the process can be monitored by TLC). The resulting mixture was then subjected to aqueous treatment (using distilled water or brine) and extracted three times with ethyl acetate. The complex organic layer was dried over anhydrous sodium sulfate, filtered, and concentrated in vacuum. The product was purified by flash chromatography using ethyl acetate and hexane as solvents.

General procedure for the synthesis of N-fused indoline

A dry 10 mL Schlenk tube containing a stirring bar was charged with N-tosylhydrazone (0.1 mmol, 1.0 equiv.) and DBU (0.15 mmol, 1.5 equiv.). After purging the flask three times under vacuum and three times with argon, 1,4-dioxane (1.0 mL, 0.1 M) was added while maintaining an argon atmosphere. The reaction was carried out under a 40w 427 nm Kessil lamp reaction apparatus for 12 h (the process can be monitored by TLC). The resulting mixture was then subjected to aqueous treatment (using distilled water or brine) and extracted three times with ethyl acetate. The complex organic layer was dried over anhydrous sodium sulfate, filtered, and concentrated in vacuum. The product was purified by flash chromatography using ethyl acetate and hexane as solvents.

Assessment of cell viability via CCK-8 assay

HT1080 cells at approximately 90% confluence were trypsinized, centrifuged at 800 × g for 3 min, resuspended, and then seeded in 96-well plates at 6×10³ cells/well. The cells were cultured for 24 h in complete DMEM (37 °C, 5% CO₂). After aspirating the medium, groups were treated as follows for 24 h: the control received 0.1 mL basal DMEM; the positive control received 0.1 mL basal DMEM with 1 μM RSL3; experimental groups received 0.1 mL basal DMEM containing 1 μM RSL3 and 10 μM indole/indoline derivatives, with six replicate wells per group. Post-incubation, the medium was replaced with 0.1 mL serum-free DMEM and 0.01 mL CCK-8 solution, followed by 2 h incubation (37 °C, 5% CO₂). Absorbance at 450 nm was measured using a microplate reader. Cell viability (%) was normalized to untreated controls to evaluate the protective effects of indole/indoline derivatives against RSL3-induced ferroptosis in HT1080 cells.

Lipid peroxidation assay

The fluorescent probe C11-BODIPY (581/591) (D3861, Thermo Fisher Scientific) was employed to monitor intracellular levels of lipid peroxidation. Cells were collected for analysis in an Agilent NovoCyte flow cytometer (2060 R) after 1 h incubation with the probe. Data were analyzed using FlowJo software (FlowJo).

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Author contributions

Q.L. conducted the synthetic experiments. R.W. and X.H performed the DFT calculations. S.D. and Z.L. synthesized a portion of the materials. H.X. and X.F. carried out the biological evaluation. Y.Z. conceived this work. Y. Z and W.D. Z directed the project. W.D. Z, X.H, X.F., and Y.Z. wrote the paper.

Peer review

Peer review information

Nature Communications thanks Xiaolong Yang and the other anonymous reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Data availability

All the data generated in this study are provided in the Supplementary Information/Source Data file. Source data are provided with this paper. All data are available from the corresponding author upon request.

Small molecule drug catalog number:

RSL3: T3646, TargetMol

IKE: S8877, Selleck Chemicals

Fer-1: S7243, Selleck Chemicals. Source data are provided with this paper.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s41467-026-71337-8.