Hexafluoroisopropanol-Mediated Cascade Cyclization for Construction of α‑Amino-γ-spirolactones

Abstract

Given the ubiquity of spirolactones in pharmaceutical agents, natural products, and small-molecule chemical probes, considerable research has been devoted to streamlining their synthesis. Yet, the preparation of α-amino-γ-spirolactones, which combine two key motifs in drug discoveryspirocycle and amine unitsremains underdeveloped. Here, we demonstrate that readily accessible building blocksamines, glyoxylic acid, and unactivated alkenescan be rapidly converted to complex three-dimensional α-amino-γ-spirolactones and related compounds employing a hexafluoroisopropanol-promoted cascade cyclization. This protocol is general, efficient, scalable, and applicable to the mid- and late-stage functionalization of medicinally relevant molecules, while displaying excellent sustainability metrics. All those features, along with preliminary results indicating the potential bioactivity of spirolactones against cancer cells, make the reported methodology a promising tool for drug discovery.

Introduction

Creating molecular complexity from simple and readily available building blocks has become one of the focal points of synthetic and medicinal chemists to access densely functionalized architectures that expand our chemical space. In particular, a significant emphasis has been placed on forging molecules with a three-dimensional structure, in which saturation and complexity are increased to improve their biological activity. − As a result, molecules incorporating spirocyclic units have sparked a growing interest in drug discovery. − This structural element brings other benefits, e.g., an increased solubility, a reduced conformational entropy, and a control of the spatial placement of the functional groups present in the molecule, making spirocyclic compounds valuable for in silico fragment-based screening. Lastly, the ability to fine-tune the structural and functional aspects within these frameworks by varying ring size, adjusting electronic properties, and calibrating substituent effects enables additional exploration of their interactions with target binding sites. In this context, spirolactones have emerged as privileged scaffolds in commercial drugs and pharmaceutical candidates, including spironolactone - the first FDA-approved spirocyclic drug -, psilostachyin, and crotonianoid A (Scheme A). − Accordingly, synthetic chemists have endeavored to develop novel approaches to access spirocyclic scaffolds in a simple fashion.

1. Importance of Spirolactones and Strategies to Access α-Amino-γ-spirolactones.

A further common feature in the context of drug discovery in small-molecule drugs is the presence of an aliphatic amine functionality or a nitrogen heterocycle. Considering both the increased interest in three-dimensional structures and the prevalence of nitrogen-containing functional groups, we wondered whether we could develop a general and efficient strategy to construct nitrogen-containing spirolactones. Such a framework is found in natural products such as speradine C and tryptoquivaline F, and could lead to new, potentially biologically relevant scaffolds. We notably focused our attention on the preparation of compounds incorporating α-amino-γ-butyrolactone motifs as they are common structural features in bioactive molecules such as N-(3-oxododecanoyl)-l-homoserine lactone (OdDHL), , and they could serve as direct precursors to biologically relevant unnatural α-amino acids, notably γ-hydroxyamino acids.

Only a few methods have been reported for accessing α-amino-γ-spirolactones, but three stood out (Scheme B). In 2000, the group of Li described a three-component tandem cyclization involving electron-deficient primary anilines, alkenes, and glyoxylates in the presence of a combination of Lewis and Brønsted acids, while, in 2023, the group of Maulide developed an elegant variant of this transformation using an aminal as a precursor in a trifluoroacetic acid (TFA)-mediated protocol. In that case, the authors also demonstrated the applicability of their approach using ethyl glyoxylate and selected amines as substrates. Nevertheless, both methods suffer from a limited substrate scope, notably with respect to the amine moiety, and the bioactivity of the synthesized compounds was not evaluated. In 2021, Hua and Wang reported an enantioselective synthesis of such spirolactones; however, this method requires engineered substrates, limiting its generality.

Whereas multicomponent reactions generally represent a streamlined process for achieving a rapid increase of molecular complexity, − they can be operationally challenging because they require combining several chemical events that are compatible with one another. Furthermore, side reactions or even process shutdown due to promoter trapping by a substrate or a reacting intermediate must be avoided. Overall, all these factors can drastically narrow the scope of such transformations. In this context, we assumed that the use of 1,1,1,3,3,3-hexafluoropropan-2-ol (HFIP) could represent a powerful tool for the development of new multicomponent reactions. Indeed, HFIP - used as a solvent or as an additive - has recently enabled previously reluctant reactions through its ability to activate various substrates, extend the lifetime of reactive intermediates, and prevent any deactivation because of its strong H-bond donating ability and low nucleophilicity/Lewis basicity. −

Here we notably imagined that a multicomponent reaction relying on HFIP could provide an efficient access to α-amino-γ-spirolactones, using the following design plan (Scheme C): HFIP would promote a fast formation of an electrophilic iminium ion (Int-I), occurring through the condensation of glyoxylic acid with an amine. This iminium would then undergo an aza-Prins-type reaction with an alkene to generate the corresponding carbocation Int-II, followed by cyclization with the carboxylic acid to yield the target spirolactone. Such a process would be particularly favored in HFIP since it is well-documented that iminium species are either stabilized by HFIP or react rapidly with HFIP to form a hexafluoroisopropyl ether that can allow a gradual release of the key cationic reacting intermediate, preventing its decay through undesired side reactions under the reaction conditions.

Herein we disclose a one-pot condensation/aza-Prins-type reaction/cyclization sequence for the assembly of a wide range of α-amino-γ-spirolactones and α-amino-γ-butyrolactones under mild conditions, resting on the inherent properties of HFIP. This protocol utilizes inexpensive feedstock chemicals - amines, glyoxylic acid, and alkenes - and is notably applicable to the mid- and late-stage functionalization of pharmacophore fragments and medicinally relevant molecules (Scheme D). Our methodology is operationally simple, does not require any metal or additives, and only forms water as a byproduct. Moreover, this protocol proved scalable and attractive in terms of sustainability metrics, yielding compounds with antitumor activity.

Results and Discussion

We initially studied the envisioned reaction sequence using dibenzylamine 1a (1.2 equiv), glyoxylic acid 2 (1.2 equiv), and methylenecyclohexane 3a (1.0 equiv) as substrates in HFIP (0.2 M) at 40 °C (Table ). Under these conditions, spirolactone 4 was obtained in 93% yield after 4 h (Entry 1). The structure of 4 was confirmed by X-ray crystallography. Remarkably, this reaction did not require the use of any catalyst or additive. Then, a survey of solvents, including other fluorinated alcohols, was performed; however, none of them delivered the target spirolactone (Entries 2–9). The only solvent that led to the product was perfluoro-tert-butanol (PFTB), but only in 18% yield (Entry 10). Decreasing the temperature or the reaction time only led to lower yields (Entries 11 and 12).

1. Optimization Studies .

Entry	Variation from standard conditions	Yield (%)
1	None	93
2	DCM instead of HFIP	<5
3	1,2-DCE instead of HFIP	<5
4	toluene instead of HFIP	<5
5	MeNO2 instead of HFIP	<5
6	iPrOH instead of HFIP	<5
7	TFE instead of HFIP	<5
8	CH3CH(CF3)OH instead of HFIP	<5
9	CH3C(CF3)2OH instead of HFIP	<5
10	PFTB instead of HFIP	18
11	25 °C instead of 40 °C	78
12	2 h instead of 4 h	71

^aStandard reaction conditions: dibenzylamine (0.24 mmol, 1.2 equiv), glyoxylic acid (0.24 mmol, 1.2 equiv), and methylenecyclohexane (0.20 mmol, 1.0 equiv) in HFIP (0.2 M) at 40 °C for 4 h.

^bIsolated yield.

All those experiments highlight that HFIP is crucial for the transformation. As HFIP was previously shown to be able to stabilize iminium species, we hypothesized this to be the main reason for the effectiveness of the reaction. To test this hypothesis, the reaction of 1a and 2 in HFIP at 40 °C was performed in the absence of methylenecyclohexane, and the reaction mixture was studied by NMR spectroscopy. After taking an aliquot and diluted it with CD₂Cl₂, ¹H NMR showed the characteristic signals of an iminium ion with a ratio iminium/dibenzylamine 3:1 (see Supporting Information for details). While the NMR signals of the iminium ion were also detected in PFTB, albeit in a lower amount than in HFIP (ratio iminium/dibenzylamine 1.1:1), it was not the case in the other solvents (or only as traces), suggesting this intermediate to be more stable in HFIP than in other solvents.

With optimized reaction conditions in hand, we investigated the scope of the reaction. We first evaluated the reactivity of various secondary amines with glyoxylic acid 2 and methylenecyclohexane 3a (Scheme ). Moderately nucleophilic amine nucleophiles such as diarylamines bearing electron-donating and -withdrawing groups at different positions were used effectively to afford the corresponding spirolactones 5-11 in 49–75% yields. The reaction also accommodated N-alkyl anilines to provide the target spirolactones 12-19, notably incorporating amino acid–based derivatives such as glycine (18, 57%) and frameworks of interest such as tetrahydro-1H-benzo­[b]­azepine (19, 53%). On the other hand, no reaction was observed in the case of N-alkyl heteroarylamine such as 4-(methylamino)­pyridine. Here, the pyridine unit likely interferes with the H-bond network of HFIP, thereby rendering it ineffective to promote the reaction. Importantly, the reaction proved chemoselective in the presence of functionalities such as an alkene (16) or alkyne (17), neither competing with the exocyclic double bond nor leading to additional reactions. Then, we examined the reactivity of N,N-dialkylamines containing a benzyl or a naphthylmethyl moiety along with a range of linear functionalized alkyl substituents, which resulted in the streamlined formation of diverse spirocycles in high yields (up to 88%, 20-27). The reaction was not affected by the presence of electron-withdrawing groups such as CN and ester (23 and 24) and was compatible with sterically hindered branched alkylamines (25, 88%). The protocol was also extended to aliphatic amines without a benzyl unit, notably 2-phenylethylamine (28, 80%) and ethanolamine (29, 96%) derivatives. In addition, our method allows the introduction of several biologically relevant cyclic amines such as azetidine (30), pyrrolidine (31), azocane (32), piperidine (33-37), morpholine (38), thiomorpholine (39), and piperazine (40) in moderate to high yields (36–89%). In the case of piperidine, the reaction was tolerant to the presence of various functional groups, including free alcohol (33), aryl (34), ester (35), trifluoromethyl (36), and cyano (37). Importantly, the acid-sensitive Boc protecting group remained intact as illustrated in 40 (89%). While the reaction is also compatible with electron-deficient aromatic primary anilines (41 and 42, 81 and 80%), its limitations are its compatibility with primary alkylamines and electron-rich anilines, for which the desired reactivity was not observed. However, this issue was circumvented due to the excellent reactivity of N,N-dialkylamines. For instance, the benzyl group of 23 was easily cleaved in the presence of ammonium formate and Pd/C to give compound 43 in 91% yield, which would formally result from the reaction with the corresponding primary aliphatic amine. In the same vein, the N,N-dibenzyl group of 4 was removed by hydrogenation with H₂ on Pd­(OH)₂/C to furnish the corresponding primary aliphatic amine 44 in 94% yield.

2. Scope of the Transformation.

Our methodology was also applicable to a broad range of alkene partners, whether in terms of ring-size or functional group tolerance, to furnish spirolactones 45-50 in high yields (up to 91%). Here naturally occurring alkenes such as camphene enabled the efficient and selective construction of the corresponding spirolactones 51 (84%), indicating the possibility of expanding the existing synthetic toolbox for chimeric molecules. Of note, no additional multicomponent reactions or competing side reactions were observed when alkenes bearing heteroatom functionalities were employed (45 and 46). Importantly, tri- and tetrasubstituted alkenes were compatible with our protocol to access compounds 52 and 53 in 87% and 85% yields, respectively. Furthermore, we could extend our protocol to other alkene partners such as isobutene (54), styrene (55 and 56), indene (57), trans-1-phenyl-1,3-butadiene (58) and a 1,3-enyne (59) to produce densely functionalized lactones in 41–87% yields along with a moderate to high control of the diastereoselectivity (up to 10:1 diastereoisomeric ratio), which likely results from t. Yet, the reaction proved ineffective with monosubstituted aliphatic alkenes such as 1-hexene. In that case, the alkene is not nucleophilic enough to react with the iminium intermediate and remained intact under the reaction conditions at 40 °C. On the other hand, heating at higher temperatures resulted in gradual degradation of the olefin.

In the case of allene substrates, the reaction notably opens the access to amine-containing furan-2-ones such as 60 (83%), whose structure was ascertained by X-ray crystallography. Another feature of our protocol is the possibility to apply it to the underdeveloped preparation of valuable bridged piperidine-γ-butyrolactones such as 61 (96%) via the use of substrate connecting alkene and amine functionalities in a two-component reaction.

To highlight the versatility of the transformation for potential drug design, we explored the combination of these three-dimensional spirolactones with complex and densely functionalized amines, including pharmacophores and biologically relevant compounds (Scheme ). For instance, several fragments of drugs, including vilazodone, quetiapine, ziprasidone, donepezil, bepotastine, and paliperidone, were successfully engaged in our process to furnish the target products 62-67 in yields ranging from 47 to 64%. Moreover, marketed drugs such as norfloxacin, amoxapine, fasudil, crizotinib, desloratadine, paroxetine, and fluoxetine, which contain a secondary alkylamine unit, smoothly reacted with 2 and 3a to deliver the corresponding spirolactones 68-74 in moderate to high yields (36–87%). These results further demonstrate the large functional group compatibility of our protocol. Not surprisingly, in most cases, no control of the diastereoselectivity was observed, as the amine substrates only contain remote stereocenters that are not proximal with the reacting iminium site. Remarkably, this transformation is not only applicable to the synthesis of complex spirolactones but also to the preparation of lactones, using styrene as an alkene component (75-80, 63–89%). In the case of the reaction between fasudil, styrene, and glyoxylic acid, we were able to determine the configuration of the major diastereoisomer by X-ray crystallography.

3. Mid- and Late-Stage Functionalization of Pharmacophores and Bioactive Molecules.

To illustrate the utility of our transformation, we evaluated the scale-up of the reaction and its sustainable metrics (Scheme ). The reaction between dibenzylamine 1a (6.0 mmol, 1.18 g, 1.2 equiv), glyoxylic acid 2 (6.0 mmol, 0.44 g, 1.2 equiv), and methylenecyclohexane 3a (5.0 mmol, 0.48 g, 1.0 equiv) in HFIP (25 mL, 0.2 M) at 40 °C for 4 h did not lead to a significant drop in yield, affording spirolactone 4 in 92% (4.6 mmol, 1.60 g). Regarding the sustainability metrics of this multicomponent reaction, we calculated it to be 95.0% atom-economical, 87.4% atom-efficient, 100% carbon-efficient, and 76.2% reaction-mass-efficient. Moreover, the reaction displays a remarkably low E-factor (<1) since we recovered a significant amount of HFIP by distillation at the end of the reaction (24.5 mL), and no workup was necessary before purification by flash column chromatography. Overall, the radar chart of our protocol highlights its excellent sustainability.

4. Scale-Up Experiment and Evaluation of Sustainability Metrics.

Having established an efficient route for the one-pot synthesis of a variety of spirolactones, we conducted a preliminary screening of their antitumor activity to test the utility of these molecules for potential medical applications. We notably evaluated representative spirolactones against two human cancer cell lines - PC-9 (lung cancer cell line) and PC-3 (prostate cancer cell line) - by using the CCK8 (Cell Counting Kit) assay, in which 5-fluorouracil (5-FU) serves as a positive control (see the Supporting Information for details). We found that most spirolactones exhibit cytotoxic activity against the tested cell lines (see Supporting Information, Table S1). Among them, compound 72 has a concentration inducing 50% cell growth inhibition (IC₅₀) that is comparable to that of 5-FU in the PC-9 cell line (7.88 ± 0.74 μM vs 9.81 ± 0.29 μM) (Scheme ). Importantly, the spirolactone motif is not innocent in the observed bioactivity, as desloratadine alone does not display any cytotoxic activity. Here the spirocycle provides three-dimensional rigidity and increases the fraction of sp³ carbon carbons (Fsp³), which may reduce interactions with metabolic enzymes such as cytochrome P450 (CYP450), thereby enhancing the metabolic stability of the corresponding compound. To determine if it is the case with our compounds, we conducted a preliminary human microsomal stability assay using compound 72. After a 100 min incubation in human microsomes, the remaining amount of 72 was higher than that of desloratadine. The calculated half-life of 72 also extended compared to desloratadine (t _1/2 = 81.30 min vs t _1/2 = 55.42 min), indicating improved metabolic tolerance likely due to the increased proportion of sp³-hybridized carbons.

5. Antiproliferative Activities of 5-FU, 72, and Desloratadine 72-S against PC-9 and Human Microsomal Metabolic Assay .

^a IC₅₀ values are the mean ± standard error (SEM) of three independent experiments.

To obtain further insights into the structural diversity of the α-amino-γ-spirolactones 4-80, we calculated their three principal moments of inertia (PMIs). , After normalization to give the mass-independent ratios NPR₁ and NPR₂, the so-called PMI descriptors, each molecule can be placed on a ternary diagram where proximity to the rod, disc, or sphere vertices quantitatively reflects its overall shape (Scheme ). The sprawl of our molecules across the PMI plot layout demonstrates the stereochemical complexity that can be realized using our methodology. Despite the intrinsic three-dimensionality of the spirolactone core, most of the substrates cluster in the rod- and disc-like regions, predominantly caused by the extended geometries of bulky substituents at nitrogen and the carbon scaffold. Yet, multiple substrates feature extended three-dimensionality by extending into the spherical region.

6. Analysis of the Molecular Shape of Species 4-80 and 5-Fluorouracil (5-FU) through a PMI Plot .

^a Each point in the blue triangle represents one compound; positions near the top-left vertex indicate rod-like shapes, near the bottom vertex disc-like, and near the top-right vertex sphere-like.

To determine if the PMI analysis could be used as a tool to predict the bioactivity of such compounds based on a hit found (72), we assessed 65 and 77, which are located in the same area as 72 (red dots in Scheme ). Remarkably, while 65 exhibits moderate bioactivity (see Supporting Information, Table S1), 77 displays a similar IC₅₀ as the ones of 5-FU and 72 (6.50 ± 0.60 μM) (Scheme ). In a similar vein as 72, the fragment of paliperidone alone does not show any cytotoxicity without the lactone motif. While those results do not allow for drawing any definitive conclusion on the utility of the PMI analysis, it is certainly a new tool to consider for potential medicinal applications.

7. Antiproliferative Activities of 5-FU, 77, and 77-S against PC-9 and Human Microsomal Metabolic Assay.

^a IC₅₀ values are the mean ± standard error (SEM) of three independent experiments.

Conclusion

In summary, we have developed a metal- and additive-free three-component reaction, relying on the solvent HFIP, to access nitrogen-containing spirolactones and related compounds from simple building blocks such as amines, glyoxylic acid, and unactivated alkenes. The structural and functional diversity exhibited by this protocol offers a versatile and flexible approach for the efficient synthesis of complex three-dimensional molecules under mild conditions. The robustness of our methodology was further demonstrated by its scalability and applicability to the mid- and late-stage functionalization of medicinally relevant molecules. Importantly, the excellent sustainable metrics of this transformation and the potential bioactivity of the compounds obtained indicate the viability of using such a process in the development of new therapeutics.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/jacsau.5c00958.

• Experimental procedures and characterization data, additional experimental details, bioactivity data, normalized principal ratios and ¹H, ¹³C, and ¹⁹F NMR spectra (PDF)

∥.

H.L. and F.H. contributed equally. The manuscript was written through contributions of all authors. All authors have given approval to the final version of the manuscript. CRediT: Haoran Li data curation, formal analysis, investigation, methodology; Fangting Huang data curation, formal analysis, investigation, methodology; Claire Miesch investigation, methodology; Robert J. Mayer investigation, software; Yuwei Liu investigation, methodology; Jianlian Huang investigation, methodology; Lei Ma data curation, formal analysis, investigation; Zhongyi Zeng data curation, formal analysis, investigation; Zhi Zhou data curation, formal analysis, investigation; David Lebœuf conceptualization, formal analysis, project administration, resources, supervision, writing - original draft, writing - review & editing; Wei Yi conceptualization, formal analysis, funding acquisition, project administration, resources, supervision, writing - review & editing; Shengdong Wang conceptualization, formal analysis, funding acquisition, project administration, resources, supervision, writing - original draft, writing - review & editing.

Guangdong Zhujiang Talent Program (0920220229), Guangdong Basic and Applied Basic Research Foundation (2024A1515011476), and National Natural Science Foundation of China (82273795).

The authors declare no competing financial interest.