Light-enabled scalable synthesis of bicyclo[1.1.1]pentane halides and their functionalizations

Abstract

In 2012, bicyclo[1.1.1]pentanes were demonstrated to be bioisosteres of the benzene ring. Here, we report a general scalable reaction between alkyl iodides and propellane that provides bicyclo[1.1.1]pentane iodides in milligram, gram and even kilogram quantities. The reaction is performed in flow and requires just light; no catalysts, initiators or additives are needed. The reaction is clean enough that, in many cases, evaporation of the reaction mixture provides products in around 90% purity that can be directly used in further transformations without any purification. Combined with the subsequent functionalization, >300 bicyclo[1.1.1]pentanes for medicinal chemistry have been prepared. So far, this is the most general and scalable approach towards functionalized bicyclo[1.1.1]pentanes.

A general and scalable approach to prepare functionalized bicyclo[1.1.1]pentanes is reported. The key reaction between alkyl iodides and propellane is performed in flow and requires just light; no catalysts, initiators or additives are needed. Combined with the subsequent functionalization, >300 bicyclo[1.1.1]pentanes for medicinal chemistry have been prepared.

Main

The benzene ring is the most popular ring in drugs^1,2 and natural products³. In 2012, bicyclo[1.1.1]pentane (BCP) was demonstrated to mimic the para-substituted benzene ring in a biologically active compound (Fig. 1a)⁴. Since then, BCPs have been playing an important role in chemistry^5–32. Synthesis and applications of BCPs are covered in at least ten recent reviews^5–14. Moreover, >300 patents describe an application of BCPs in drug discovery projects (Fig. 1b). Most of these compounds bear a (hetero)aromatic substituent, the hydrogen atom or a carboxylic group derivative at the bridgehead position of the BCP core, however, alkyl substituents are rare.

Fig. 1. Application and synthesis of BCPs.

a, The concept begins with BCPs as saturated bioisosteres of the para-substituted benzene ring^3,4. b, BCPs in drug discovery projects. In the first two molecules (Pfizer in 2012, ref. ⁴ and GlaxoSmithKline (GSK) in 2017, ref. ¹⁶) BCPs have been used as saturated benzene bioisosteres. c, Known approaches to alkyl-substituted BCP iodides. r.t., room temperature; fac-Ir(ppy)₃, fac-tris(2-phenylpyridine)iridium(III); FG, functional group.

Worth specific mentioning is a recent collaboration between Pfizer and the Baran laboratory on developing a scalable ‘strain-release’ amination of propellane^33,34. This study allowed the preparation of BCP amines with no substituents at the bridgehead position.

Aliphatic substituents increase the fraction of sp³-hybridized carbon atoms (F(sp³)) in bioactive molecules and, therefore, it is not surprising that medicinal chemists favour using them nowadays^35,36. In this context, alkyl-substituted BCPs are conceptually interesting, yet almost unknown.

In this work, we have developed a general scalable reaction between inexpensive starting materials—alkyl iodides and propellane—that gives alkyl-substituted BCP iodides in milligram, gram and even kilogram quantities. The reaction proceeds in flow and requires only light. No catalysts, initiators or additives are needed. The transformation is so clean that in many cases, evaporation of the reaction mixture provides products in around 90% purity that can be directly used in the next steps without any purification. The subsequent modifications of the obtained products allowed the preparation of >300 BCP building blocks for use in medicinal chemistry. So far, this is the most general and scalable approach towards functionalized BCPs.

Results

Optimization

In a search for a general scalable method towards alkyl-substituted BCPs, we focused our attention on the reaction of alkyl iodides with propellane. In 1991, this reaction was shown to take place under irradiation with a broad-wavelength Hanovia mercury lamp in a Pyrex vessel (Fig. 1c)³⁷. In 2000, it was demonstrated that the addition of an equimolar amount of methyl lithium also promoted the reaction (Fig. 1c)³⁸. The challenges associated with the ultraviolet irradiation in Pyrex glassware and the low compatibility of methyl lithium with various functional groups lowered the practical potential of both methods^39–43. In 2018, scientists discovered that triethylborane initiated the reaction leading to the formation of products in good yields (Fig. 1c)^44–46. In 2019, the reaction scope was improved by performing the reaction in the presence of the fac-Ir(ppy)₃ catalyst under the photoredox conditions^47,48. The key disadvantage of the last method was the relatively high price of the metal catalyst (price of fac-Ir(ppy)₃ (Aldrich): €852 for 250 mg). It is worth noting that some activated alkyl iodides, such as CF₃I, HCF₂I and EtO₂CCF₂I, were reported to react with propellane at room temperature without any initiation or catalysis^49–51. The reaction was slow, and typically took place over 48–72 hours.

Known approaches towards alkyl-substituted BCPs are depicted in Fig. 1с. These reactions were described on a milligram scale. Therefore, we first tried these protocols on a gram scale with the most challenging alkyl substrate: methyl iodide (MeI). The methyl radical is the most unstable among all common alkyl radicals. We thought that if we could elaborate on a scalable protocol for the reaction of MeI, it would work with other alkyl iodides too.

The reaction of MeI with propellane (1.2 eq.) in the presence of methyl lithium³⁷ or iodine gave only traces of the needed product 1 (Table 1, entries 1 and 2). Initiation with BEt₃ (ref. ⁴⁴) gave the product 1 in 31% yield, however, an extensive formation of polymeric products was observed (entry 3). Catalysis with fac-Ir(ppy)₃ (refs. ^45,46) led to the formation of a complex mixture (entries 4 and 5). With an excess of propellane (2 eq.), however, we obtained iodide 1 in 14% yield. Next, we attempted the reaction under various photochemical conditions in batch with no catalysts/initiators (entries 7–10). Irradiation of the reaction mixture at 254 or 310 nm gave <20% of the needed product (entries 7 and 8). Irradiation at 450 nm (blue light-emitting diode (LED)) did not promote the reaction (entry 9). However, the irradiation at 365 nm in batch allowed obtaining the desired product 1 in 43% yield (entry 10). After further optimization, we found that performing the reaction in flow for 30 minutes allowed increasing the yield to 62% (entry 11).

Table 1.

Optimization of synthesis of BCP 1

Entry	Conditions	Yield (%)b
1	MeLi (1 eq.), r.t., 24 h	5
2	I2 (0.25% mol.), r.t., 24 h	16
3	BEt3 (10% mol.), Et2O, r.t., 24 h	31
4	fac-Ir(ppy)3 (2.5% mol.), Et2O, 450 nm, 12 h	Polymerization
5	fac-Ir(ppy)3 (2.5% mol.), tBuCN, 450 nm, 12 h	Polymerization
6	fac-Ir(ppy)3 (2.5% mol.), tBuCN, 450 nm, 12 h (2 eq. propellane)	14
7	254 nm, Et2O, r.t., 24 h, in batch	12
8	310 nm, Et2O, r.t., 24 h, in batch	17
9	450 nm, Et2O, r.t., 24 h, in batch	<5
10	365 nm, Et2O, r.t., 24 h, in batch	43
11	365 nm, Et2O, r.t., 30 min, in flow	62
12	365 nm, Et2O, r.t., 30 min, in flow (2 eq. MeI)	73
13	365 nm, MeOtBu, r.t., 30 min, in flow	51
14	r.t., 30 min (control)	ND
15	r.t., 24 h (control)	6c
16	365 nm, Et2O, r.t., in flow (855 g of product)	72c

Scale, 10 g of MeI in each experiment. ND, not detected.

^aSolution of propellane (0.7 M) in Et₂O-CH₂(OEt)₂.

^bIsolated yield. Distillation as a purification method.

^cCrystallization as a purification method.

It is worth noting that all experiments in entries 1–11 (Table 1) were performed under the standard conditions: MeI (10 g; 1 eq.) and propellane (1.2 eq.). In each case, the product was isolated by distillation under reduced pressure. Moreover, performing the reaction with an excess of MeI (2 eq.) led to an improvement of the yield to 73% (entry 12). However, because many alkyl iodides are expensive, for further studies we used the reaction conditions that require an almost equimolar amount of reagents (entry 11).

It is important to mention that propellane can also be synthesized in MeO^tBu instead of Et₂O (also Supplementary Information, page 27). Its reaction with MeI under the developed conditions also worked and provided product 1 in 51% yield (entry 13).

Control experiments revealed that without irradiation the reaction of MeI with propellane did not proceed efficiently (entries 14, 15).

Having an optimized protocol in hand (Table 1, entry 11), we synthesized pure BCP 1 in 855 g amount in one run with almost no additional modifications (Table 1, entry 16; also Supplementary Information, pages 62–63). In this case, however, we isolated the product (72% yield) from the reaction mixture by a low-temperature crystallization from pentane.

Scope

Next, we studied the generality of the developed method. First, we tried other primary alkyl iodides (Fig. 2). Given the rise of deuterated compounds in modern drug discovery^52,53, we performed an addition of CD₃I to propellane under standard conditions to obtain product D₃-1 in 68% yield. The reaction worked well with other alkyl iodides (3–7), oxetane-containing substrates (8–10), tetrahydrofuran (11) and tetrahydropyran-containing molecules (12, 13). N-Boc-protected azetidines (14), pyrrolidines (15) and piperidines (16, 17) performed equally well in the reaction. In addition, Bpin (18), PO(OEt)₂ (19) and ^tBuC(O)O groups (20) were compatible with the reaction conditions. Given the importance of organofluorine compounds in modern medicinal chemistry^54–56, we performed the reaction with various fluorinated alkyl iodides to obtain BCPs 21–27 in 69–92% yield. The structure of products 14 and 19 was confirmed by X-ray analysis.

Fig. 2. Synthesis of BCPs from primary alkyl iodides.

Reaction conditions were as follows: a solution of alkyl iodide (1 eq.) and propellane (0.7 M in Et₂O/CH₂(OEt)₂; 1.2 eq.) in diethyl ether was passed through a coil (irradiated area 160 ml) with a flow rate 10 ml min⁻¹ under irradiation with 365 nm LED (radiated power 420 W). Residence time was 16 min. Typical scale was 5–10 g of alkyl iodide. Compounds 1, 2 and 22 were additionally obtained on a multigram scale. Synthesis of compounds 6, 40, 42–47 was performed on 0.1–2 g scale (irradiated coil 7.6 ml; flow rate 0.75 ml min⁻¹; irradiation 365 nm LED; radiated power 257 W and residence time 10.1 min). [MeLi]: MeLi (1 eq.), CH₂(OEt)₂, 24 h, r.t., in batch. [BEt₃]: BEt₃ (0.1 eq.), Et₂O, 24 h, r.t., in batch; [Ir]. fac-Ir(ppy)₃ (2.5% mol.), ^tBuCN, 12 h, r.t., in batch. X-ray crystal structure of compounds 14 and 19. Hydrogen atoms are omitted for clarity. Boc, tert-butoxycarbonyl protecting group.

Various functional groups, such as nitrile (28–30), ester (31–33), active chlorine (34) and bromine atoms (35), alcohol (36, 37) and NHBoc (38–40) were compatible with the reaction conditions. Diverse cores including 2-oxabicyclo[2.1.1]hexane (44)^57–59 and oxa-spirocycles (45)⁶⁰ also gave the desired BCP iodides 41–47 in 41–69% yield.

We also studied the behaviour of secondary alkyl iodides (Fig. 3). The protocol efficiently worked for isopropyl (48), isobutyl (49) and cycloalkyl (50–54) iodides.

Fig. 3. Synthesis of BCP halides and bicyclo[3.1.1]heptanes.

Reaction conditions were as follows: a solution of alkyl iodide (1 eq.) and propellane (0.7 M in Et₂O/CH₂(OEt)₂; 1.2 eq.) in diethyl ether was passed through a coil (irradiated area 160 ml) with a flow rate 10 ml min⁻¹ under irradiation with 365 nm LED (radiated power 420 W). Residence time was 16 min. Typical scale 5–10 g of alkyl iodide. Compounds 48, 60 and 63 were also obtained on a multigram scale. [MeLi]: MeLi (1 eq.), CH₂(OEt)₂, 24 h, r.t., in batch; [BEt₃]: BEt₃ (0.1 eq.), Et₂O, 24 h, r.t., in batch; [Ir]: fac-Ir(ppy)₃ (2.5% mol.), ^tBuCN, 15 h, r.t., in batch. X-ray crystal structure of compound 55. Hydrogen atoms are omitted for clarity.

Four-to-six-membered rings with oxygen (55–58), sulfur (59) and N-Boc (60–64) gave the desired products in 30–92% yield. Secondary, MeCH(I)CO₂^tBu, and tertiary, Me₂C(I)CO₂^tBu, iodides also reacted with propellane to provide products 65, 66 in lower yields of 25–36% due to a problematic purification⁶¹. Various fluoroalkyl iodides (67–73) and even bromides (74–78)⁶² were compatible with the reaction conditions too. The structure of product 55 was confirmed by X-ray analysis.

Several representative (hetero)aromatic iodides also were subjected to the standard reaction conditions, and products 79–81 were obtained in 47–90% yield. Phenyl iodide did not react, however.

Recently, bicyclo[3.1.1]heptanes were proposed to mimic the meta-substituted benzene ring in bioactive compounds⁶³. In this context, we studied the reaction of [3.1.1]-propellane with two representative alkyl iodides under the above-developed conditions. The desired bicyclo[3.1.1]heptanes 82 (45%) and 83 (38%) were obtained as a result of these efforts (Fig. 3).

Scalability

Most of the syntheses depicted in Figs. 2 and 3 were performed with 5–10 g of starting alkyl iodides. The typical reaction time was less than 30 minutes. Only syntheses of BCPs 6, 40, 42–47 were performed on a smaller scale due to the low availability of the corresponding alkyl iodides.

For many examined substrates, we compared the performance of our conditions with the literature protocols on the same scale (Figs. 2 and 3). MeLi gave poor yields of the desired products bearing functional groups; initiation with triethylborane often gave good results (3–5, 55, 57, 60 and so on). In each case, however, a standard aqueous workup followed by purification by column chromatography or distillation was needed.

The photochemical protocol developed here gave the best yields of products in all cases, where different protocols were examined and compared. In many cases, the reaction was so clean that evaporation of the reaction mixture provided products with around 90% purity that can be directly used in the next steps without any purification. It allowed us therefore to subsequently scale up the preparation of BCP iodides 1–3, 5, 7, 14–16, 21–27, 48, 50, 55, 57, 59–63, 67–69, 71, 75 and 77 to 50–800 g quantities (also Supplementary Information, pages 62–70).

Mechanism

Product 62 (Fig. 3) was obtained from the derivative of the optically pure (2S,4S)-4-iodoproline as a 3:2 mixture of two diastereomers at C(4)-atom (Fig. 3). This observation suggested the radical mechanism of the reaction with the initial photochemical formation of the configurationally unstable alkyl radicals. To validate this hypothesis, we performed ‘radical clock’ experiments (also Supplementary Information, pages 71–87)⁶⁴. Alkyl iodide 84 was reacted with propellane under the developed conditions to selectively form the ring-opened alkene 85 (Fig. 4). In the nuclear magnetic resonance (NMR) spectroscopy of the crude reaction mixture, we did not observe even traces of the cyclopropane ring. In the presence of (2,2,6,6-tetramethylpiperidin-1-yl)oxyl (TEMPO), the formation of product 85 was not observed. In the absence of propellane, an isomerization of iodide 84 into compound 86 slowly took place. Similarly, the reaction of alkyl iodide 87 with propellane gave mostly the rearranged cyclopentane-containing product 88 with only traces of alkene 89 (88:89 = 5:1, Fig. 4). These experiments supported the original hypothesis of the radical pathway of the reaction.

Fig. 4. Radical clock experiments with alkyl iodides 84 and 87.

^aNMR yield using 1,3,5-trimethoxybenzene as an internal standard.

Modifications

Having a practical and scalable protocol towards alkyl BCP iodides in hand, we converted them into various BCP-containing building blocks (compounds with one or two functional groups) for use in medicinal chemistry. Treatment of BCP iodides with t-BuLi in Et₂O followed by trapping of the formed carbanions with (ⁱPrO)Bpin gave boron pinacolates a (Fig. 5)^65–73. The reaction of the last with potassium fluoride in an acetone–water mixture smoothly gave trifluoroborates b. Oxidation of boron pinacolates with H₂O₂ gave alcohols c (ref. ⁷⁴). Trapping of BCP-carbanions with diverse electrophiles was also studied. Reaction with BocN=N-Boc followed by the acidic N-Boc deprotection produced hydrazines d (Fig. 5)^75,76. Reaction with sulfur dioxide followed by the oxidative chlorination of the intermediate sulfinate salts gave sulfonyl chlorides e (refs. ^77–81). Reaction with hexachloroethane afforded BCP chlorides f. An analogous reaction with 1,1,2,2-tetrabromo-1,2-difluoroethane provided BCP-bromides g. The addition of CD₃OD followed by the N-Boc deprotection gave deuterated amines 60h and 63h. The addition of ethyl formate gave aldehydes i. Treatment of carbanions with methanol followed by optional hydrolysis of the ester group or the N-Boc deprotection gave mono-substituted BCPs: carboxylic acids, amines and alcohols j. Reaction with dry ice gave carboxylic acids k (refs. ^82–101). Standard Curtius reactions of the last gave amines l (refs. ^83,102–106).

Fig. 5. Modifications of BCP and bicyclo[3.1.1]heptane iodides.

Reaction conditions for all conditions are shown. a, (ⁱPrO)Bpin, t-BuLi, Et₂O, −100 °C. b, KHF₂, acetone–water, r.t. c, KH₂PO₄, H₂O₂, THF–water, r.t. d, (i) BocN=N-Boc, t-BuLi, Et₂O, −100 °C; (ii) dioxane-HCl, r.t. e, (i) t-BuLi, SO₂, Et₂O; (ii) Cl₂, CH₂Cl₂–H₂O, 0–5 °C. f, (i) t-BuLi, Cl₃CCCl₃, Et₂O, −100 °C; (ii) deprotection: HCl–dioxane or TFA–CH₂Cl₂. g, (i) t-BuLi, C₂Br₂F₄, Et₂O, −100 °C; (ii) deprotection: HCl–dioxane or TFA–CH₂Cl₂. h, (i) t-BuLi, CD₃OD, Et₂O, −100 °C; (ii) deprotection: HCl–dioxane or TFA–CH₂Cl₂. i, t-BuLi, ethyl formate, Et₂O, −100 °C. j, Four methods for the reduction of the C-I bond: t-BuLi, MeOH, Et₂O, −100 °C; Raney-Ni, EtOH, ethylenediamine, r.t.; Bu₃SnH, AIBN, CCl₄, r.t.; or Pd–C, H₂, NEt₃, MeOH, r.t. k, t-BuLi, CO₂, Et₂O, −80 °C. l, (i) (PhO)₂P(O)N₃, Et₃N, t-BuOH, 95 °C; (ii) HCl–dioxane, Et₂O, r.t. Bpin, 4,4,5,5-tetramethyl-1,3,2-dioxaborolane; TFA, trifluoroacetic acid; acac, acetylacetonate; AIBN, azobisisobutyronitrile; Cu(TMHD)₂, copper bis(2,2,6,6-tetramethyl-3,5-heptanedionate); TMEDA, tetramethylethylenediamine.

The obtained BCP iodides were also compatible with radical cross-couplings. Several successful representative [Fe]- and [Cu]-catalysed reactions^107,108 of iodide 1 with ArMgCl and N-azoles were performed to obtain products 1m and 1n (Fig. 5; see also Supplementary Information, pages 175–178 for other examples).

Using this strategy, we have prepared >200 functionalized BCPs in gram quantities (Fig. 5). So far, this is the most general and scalable approach to functionalized BCPs. Many of these molecules have already found an application in drug discovery projects (Supplementary Information, pages 182–187).

Replacement of benzene by BCP in drugs

To independently validate the BCP scaffold as a saturated benzene bioisostere^5–8, we aimed to incorporate it into existing drugs. We also planned to study the impact of such replacement at the experimental physicochemical properties and biological activity. We chose the FDA-approved local anaesthetic drug Benzocaine and the antihistamine drug Buclizine with the para-substituted benzene rings.

The synthesis of a saturated analogue of Benzocaine commenced from N-Boc amino acid 90. Acidic N-Boc cleavage in ethanol and the simultaneous esterification of the carboxyl group gave the desired compound 91 as a hydrochloride salt (Fig. 6).

Fig. 6. Synthesis, physicochemical properties and biological activity of Benzocaine analogue.

a, Synthesis of saturated analogue of the local anaesthetic drug, Benzocaine, compound 91. Sol., the experimental kinetic solubility in phosphate-buffered saline, pH 7.4 (µM). logD (7.4), the experimental distribution coefficient in n-octanol–phosphate-buffered saline, pH 7.4. clogP, the calculated lipophilicity. CL_int, the experimental metabolic stability in human liver microsomes (µl min⁻¹ mg⁻¹). t_1/2 (min), the experimental half-time of a metabolic decomposition in human liver microsomes. b, Time course of the antinociceptive effect of Benzocaine and its analogue 91 in tail flick test. The data Are presented as mean ± s.e.m. (n = 5). *P(t) < 0.05 and ***P(t) < 0.001 compared with the control group (vehicle) data were analysed using a two-sided Student’s t-test without multiple comparisons. c, The area under the curve (AUC) of the withdrawal latency of Benzocaine and its analogue 91 in a tail flick test¹¹⁰. The data were presented as mean ± s.e.m. (n = 5). *P(t) = 0.026; **P(t) = 0.002 compared with the control group (vehicle) data were analysed using a two-sided Student’s t-test without multiple comparisons. n, sample size.

Source data

An impact of the replacement of the benzene ring in Benzocaine with BCP at the experimental physicochemical properties—water solubility, lipophilicity—and metabolic stability was investigated (Fig. 6). Such replacement slightly decreased the water solubility: 385 μM (Benzocaine) versus 319 (91). To estimate the influence of the replacement on lipophilicity, we used two parameters: calculated (clogP) (clogP was calculated with ChemAxon (v.22.13)) and experimental (logD) lipophilicities. According to both indices, the replacement of the benzene ring with BCP notably decreased the lipophilicity by ≥1.5 clogP/logD units. The replacement also decreased the metabolic stability, CL_int (mg min⁻¹ μl⁻¹): 83 (Benzocaine) versus 140 (91).

We also measured the experimental anaesthetic activity of Benzocaine and its analogue 91 in vivo. We studied the antinociceptive effect of both compounds using the ‘tail flick test’¹⁰⁹ in 2-month-old CD-1 female mice (Fig. 6 and Supplementary Information, pages 880–885. Study design, animal selection, handling and treatment were in accordance with Bienta Animal Care and Use Guidelines, and European Union directive 2010/63/EU). On the one hand, compound 91 was found to be less active compared to the original drug Benzocaine: no substantial difference in response time to tail flick was present throughout the observation period. On the other hand, analogue 91 demonstrated a clear analgesic activity: a notable increase in coverage of analgesia by time (area under the curve level) compared to that of the vehicle (Fig. 6).

The synthesis of a saturated analogue of Buclizine was performed from the carboxylic acid 92. Amide coupling of the latter with the appropriately N-substituted piperazine provided compound 93. Reduction of the amide group with LiAlH₄ followed by addition of hydrochloric acid gave the desired compound 94 as a hydrochloride salt (Fig. 7).

Fig. 7. Synthesis and lipid droplets fluorescent imaging assay of Buclizine analogue.

a, Synthesis of saturated analogue of the antihistamine drug Buclizine, compound 94. Sol., the experimental kinetic solubility in phosphate-buffered saline, pH 7.4 (µM). logD (7.4): the experimental distribution coefficient in n-octanol–phosphate-buffered saline, pH 7.4. clogP, the calculated lipophilicity. CL_int, the experimental metabolic stability in human liver microsomes (µl min⁻¹ mg⁻¹). t_1/2 (min), the experimental half-time of a metabolic decomposition in human liver microsomes. ^aParameter should be considered as approximate due to the high stability of compounds. b, Confocal images of the lipid droplet formation in MCF-7 cells on incubation with Buclizine and analogue 94 for 72 h. Nuclei were stained with Hoechst 33342 (cyan), lipid droplets were stained with Nile Red (red). Scale bars, 20 µm. Effectiveness of inhibition of the growth of the human cancer cell line MCF-7 (IC₅₀ index); and lipid droplet formation (EC₅₀ index) by Buclizine and its analogue 94. HATU, 1-(bis(dimethylamino)methylene)-1H-1,2,3-triazolo(4,5-b)pyridinium 3-oxid hexafluorophosphate; DMSO, dimethyl sulfoxide.

Source data

Replacement of the benzene ring in Buclizine by BCP (94) did not affect its water solubility, as both compounds were poorly soluble in water: ≤1 µM (Fig. 7). The replacement substantially decreased the lipophilicity, however, by 1–2 clogP/logD units. An impact of the replacement on the metabolic stability was not observed, as both compounds had high stability, thus outside the reliable range for measurements (Fig. 7).

Buclizine is an antihistamine agent used as a drug for the treatment of allergy symptoms and the prevention of nausea and vomiting. Recently, Buclizine was suggested for repurposing for cancer treatment, following an observation that the original target (histamine-releasing factor) and the suggested one (translationally controlled tumour protein) were identical¹¹⁰. Subsequently, Buclizine was found to exhibit cytostatic effect in MCF-7 human cancer cell line. The cell growth arrest was observed in a suppression of cell respiration followed by the resazurin reduction assay. Buclizine also induced cell differentiation, which was seen in an accumulation of intracellular lipid droplets. In this work, we tested analogue 94, for its ability to arrest cell growth and induce lipid droplets and compared it to the parent Buclizine molecule (for details, see Supplementary Information, pages 886–889). By doing so, we expected to characterize indirectly the interaction of the compounds with the tumour protein depending on the presence of an isostere in the molecule.

In the resazurin reduction assay, the original drug, Buclizine, showed moderate effectiveness (half-maximum inhibitory concentration (IC₅₀) 31.3 ± 7.8 μΜ; Fig. 7). The BCP analogue 94 behaved similarly (IC₅₀ = 24.5 ± 1.5 μΜ). In an experiment assisted by fluorescence imaging, Buclizine (half-maximum effective concentration (EC₅₀) 19 μΜ) and the BCP analogue 94 (EC₅₀ = 16 μΜ) also showed a similar onset of lipid droplet formation.

This overall preservation of activity in both analogues 91 (Benzocaine) and 94 (Buclizine) demonstrates that the benzene-to-BCP replacement is bioisosteric, thus supporting the literature data^4–8.

Summary

In 2012, BCPs were demonstrated to mimic the benzene ring in bioactive compounds⁴. Here, we report a general scalable reaction between alkyl iodides and propellane that gives alkyl-substituted BCP iodides in milligram, gram and even kilogram quantities. The reaction proceeds in flow and requires only light. No catalysts, initiators or additives are needed. The reaction is so clean that in many cases, evaporation of the reaction mixture provides products in around 90% purity that can be directly used in the next step without any purification. With the subsequent modifications, we have prepared >300 of BCPs for use in medicinal chemistry. So far, this is the most general and scalable approach towards functionalized BCPs.

We hope that this work will help process chemists at pharmaceutical companies with the preparation of bioactive BCPs suggested by medicinal chemists for clinical trials.

Methods

Synthesis of 3-iodo-BCPs and 3-iodobicyclo[3.1.1]heptanes

General protocol A

To a solution of MeI (10.00 g, 0.0704 mol, 1.00 equiv.) in Et₂O (100 ml), was added propellane (120 ml, 0.0840 mol, 0.7 M solution in Et₂O–diethoxymethane; 1.20 equiv.) under an argon atmosphere. The resulting mixture was passed through a photoreactor. The flow rate was 10 ml min⁻¹; the irradiated coil was 160 ml, irradiation was 365 nm, LED, and the irradiated power was 420 W (50% of the maximal). Before entering the irradiated area, the solution was precooled to 0 °C with Huber Unistat 510 chiller. The irradiated coil was also cooled to 0 °C with Huber Unistat 510 chiller. The temperature of the reaction mixture after the coil was around 10 °C. After passing through the coil, after around 40 min (residence time in the irradiated coil, 160/10 approximately 16 min) the solution was evaporation under a reduced pressure (around 50 mmHg, 20 °C in an external water-cooling bath), and the residue was purified by distillation under a reduced pressure (boiling point 38–40 °C at 10 mmHg). After cooling down to room temperature, the product 1 (9.21 g, 0.0438 mol, yield 62%) slowly solidified. Alternatively, the crude residue (after evaporation of the reaction mixture after the irradiation) could be purified by adding pentane (around 50 ml) and cooling the formed suspension with an external dry ice bath to around −60 °C. The formed precipitate was quickly filtered off. The filtered solid was washed with the precooled pentane (−60 °C, 50 ml) on the filter, and was dried under vacuum (20 mmHg) during 30 min at room temperature. Yield was 9.77 g, 0.0465 mol, 66%, white crystals, melting point <30 °C.

NMR spectra were analysed with MestreNova (v.11.0.3-18688).

Reporting summary

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

Author contributions

V.R., V.S., S.Z. and P.K.M. designed the experiments. V.R., V.S., V.L., S.H., D.V., I.K., D.K., S.V., Y.D., A.T., I.S., I.P., K.H., V.K. and Y.N. conducted and analysed the experiments described in this report. P.K.M. and I.S. prepared this manuscript for publication.

Peer review

Peer review information

Nature Synthesis thanks Murugaiah Subbaia and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. Primary Handling Editor: Peter Seavill, in collaboration with the Nature Synthesis team.

Data availability

The data underlying this study are available in the published article and its Supplementary Information, including experimental procedures, calculations, characterization data, copies of ¹H, ¹⁹F, ¹³C NMR spectra. Crystallographic data for the structures reported in this Article have been deposited at the Cambridge Crystallographic Data Centre, under deposition numbers CCDC 2260408 (14), 2244857 (19), 2244859 (55), 2237112 (2k) and 2244856 (21k). Copies of the data can be obtained free of charge via https://www.ccdc.cam.ac.uk/structures/.

Competing interests

The authors declare the following competing interests: V.R., V.S., V.L., S.H., D.V., I.K., D.K., S.V., S.Z., Y.D., A.T., I.S. and P.K.M. are employees of a chemical supplier, Enamine. The remaining authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s44160-024-00637-y.