2-Oxabicyclo[2.1.1]hexanes as saturated bioisosteres of the ortho-substituted phenyl ring

The ortho-substituted phenyl ring is a basic structural element in chemistry. It is found in more than three hundred drugs and agrochemicals. During the past decade, scientists have tried to replace the phenyl ring in bioactive compounds with saturated bioisosteres to obtain novel patentable structures. However, most of the research in this area has been devoted to the replacement of the para-substituted phenyl ring. Here we have developed saturated bioisosteres of the ortho-substituted phenyl ring with improved physicochemical properties: 2-oxabicyclo[2.1.1]hexanes. Crystallographic analysis revealed that these structures and the ortho-substituted phenyl ring indeed have similar geometric properties. Replacement of the phenyl ring in marketed agrochemicals fluxapyroxad (BASF) and boscalid (BASF) with 2-oxabicyclo[2.1.1]hexanes dramatically improved their water solubility, reduced lipophilicity and most importantly retained bioactivity. This work suggests an opportunity for chemists to replace the ortho-substituted phenyl ring in bioactive compounds with saturated bioisosteres in medicinal chemistry and agrochemistry.

The phenyl ring is a basic structural element in chemistry. Moreover, it is one of the most common rings in bioactive compounds 1 . However, organic compounds with more than two phenyl rings often have poor solubility and low metabolic stability-undesired effects in medicinal chemistry 2 . In this context, during the last decade, the concept 'escape from flatland' became popular 3,4 . Today, medicinal chemists prefer using F(sp 3 )-rich structures in drug discovery projects [5][6][7][8][9][10] . The replacement of the phenyl ring in bioactive compounds with saturated bioisosteres has become a popular tactic to obtain novel structures with an improved physicochemical profile [11][12][13][14] . However, most of the research in this area is devoted to the replacement of monosubstituted and para-disubstituted phenyl rings 12, .

Scaled-up synthesis
The optimized synthetic protocol is shown in Fig. 2a. It was important to identify a method that employed only available and cheap starting materials. The synthesis started from propargyl alcohol (2). Copper-catalysed reaction with phenyl magnesium bromide gave alcohol 3 in 71% yield following the reported procedure 65 . A Michael addition of the latter with methyl propiolate (4) in the presence of (1,4-diazabicyclo[2.2.2]octane) (DABCO) provided the needed diene 1. We mentioned that compound 1 partially decomposed during column chromatography and under storage at room temperature. Therefore, we decided to generate crude diene 1 in situ and use it directly in the photochemical step (Supplementary Information, page 6). A mixture of isomers 1a was obtained. After extensive experimentation, we found a way to avoid column chromatography and not lose the yield. The crude reaction mixture after irradiation (isomers 1a and benzophenone) was saponified with sodium hydroxide. A standard workup (removal of benzophenone) followed by crystallization from a hexane-MeOtBu mixture to remove the minor isomer allowed the isolation of pure major isomer 1b at 71% yield in three steps from alcohol 3. Product 1b was obtained on a ten gram scale with no column purifications.

Scope
Next, we studied the scope of the developed method. The photocyclization method tolerated various substituents on the aromatic core (Table 2). Among them were alkyl groups (5a-8a), fluorine atoms (9a-11a) and chlorine atoms (12a and 13a), methoxy groups (14a-16a) and trifluoromethyl groups (17a-19a). The reaction was also compatible with various substituted pyridines (20a-24a). In all cases, we isolated analytical quantities of intermediate esters 5a-24a by column chromatography to characterize them. On a gram scale, however, we directly used crude reaction mixtures with 5a-24a after photocyclization in the subsequent saponification step. In half of all cases, we could obtain the final

Design
In the design of a core with a similar structure to bicyclo[1.1.1]pentanes and bicyclo[2.1.1]hexanes, but having reduced lipophilicity and enhanced water solubility, we decided to insert an oxygen atom. Replacing the methylene group in bicyclo[1.1.1]pentane with the oxygen atom leads to a strained oxetane structure (Fig. 1c), which is interesting but could be labile due to possible ring opening with nucleophiles 62 . Analogous replacement in bicyclo[2.1.1]hexanes, however, gives the substituted tetrahydrofuran (Fig. 1c). That core should be more chemically stable, so we decided to make it. From a medicinal chemistry perspective, having the ether oxygen atom in the molecule is also useful, since it could serve as an additional binding site to a receptor.
Inspiration came from our previous work, where we synthesized bridgehead disubstituted 2-oxabicyclo[2.1.1]hexanes and believed they could mimic the meta-disubstituted phenyl ring in bioactive compounds 63 . This hypothesis, however, was not validated. Therefore, here we decided to prepare the disubstituted 2-oxabicyclo[2.1.1]hexanes and biologically validate them as bioisosteres of the ortho-disubstituted phenyl ring.

Synthesis
The photochemical [2 + 2] cycloaddition between alkenes proved to be a powerful strategy to construct cyclobutanes 64 . In this context, we wondered if diene 1 (easily obtained from the commercially available starting materials; Fig. 2a) would undergo an intramolecular cyclization into the needed 2-oxabicyclo[2.1.1]hexane core. Direct irradiation of diene 1 in acetonitrile under different wavelengths gave only traces of product (Table 1, entries 1-4). Irradiation with a Hanovia broad wavelength mercury lamp gave the needed product along with many side products (entry 5). Next, we tried the addition of available organic ketones for the triplet sensitization of the styrene moiety. Indeed, smooth formation of the needed product 1a (d.r. = ~4:1) was observed under irradiation at 368 nm. The best result was obtained with benzophenone (entry 7), whereas acetophenone and substituted benzophenones also worked but provided lower yields of the end product (entries 6, 8 and 9). Among all tested solvents (entries 10-13), the best outcome was obtained in acetonitrile. Without irradiation, the reaction did not take place at room temperature or with heating (entries 14 and 15).   Article https://doi.org/10.1038/s41557-023-01222-0 carboxylic acids by simple crystallization of crude reaction mixtures from various solvents ( Table 2). In the other half of the cases, column chromatography was still needed. The structure of carboxylic acids 5b and 9b was confirmed by X-ray crystallographic analysis (Fig. 2b).

Chemical stability
We also checked the chemical stability of three representative carboxylic acids, 1b, 19b and 22b (Table 2), because we suspected that some of them could decompose via a retro-Michael-type reaction. Treatment of them with aqueous 1 M hydrochloric acid or aqueous 1 M sodium hydroxide at room temperature for one day did not lead to any decomposition. All products were crystalline solids, and we stored all of them in closed vials at room temperature on the shelf. The 1 H-NMR, liquid chromatography-mass spectrometry (LC-MS) inspection after three months did not reveal any decomposition.

Incorporation into bioactive compounds
The incorporation of the 2-oxabicyclo[2.1.1]hexane scaffold into bioactive compounds was attempted next. We chose four bioactive products with the ortho-substituted phenyl ring: agrochemical fungicides fluxapyroxad and boscalid, antibacterial agent phthalylsulfathiazole and lipid-lowering agent lomitapide (Fig. 3).
Synthesis of the saturated analogue of fluxapyroxad was undertaken from carboxylic acid 11b (Fig. 3a). The standard Curtius reaction followed by acylation of the intermediate amine with the substituted pyrazole carboxylic acid gave the needed compound 29. Using an analogous tactic, compound 31-a saturated analogue of boscalid-was also obtained from carboxylic acid 17b (Fig. 3b). The saturated analogue of phthalylsulfathiazole was obtained by converting carboxylic acid 16b first into the methyl ester followed by the oxidation of the phenyl ring. Amide coupling of the formed acid with the para-substituted aniline followed by saponification of the methyl ester gave the final compound 33 (Fig. 3c). Amide coupling of carboxylic acid 19b with the correspondingly N-substituted 4-aminopiperidine gave compound 35, a saturated analogue of lomitapide (Fig. 3d).

Physicochemical parameters
In the next step, we studied the effect of the replacement of the ortho-phenyl ring by 2-oxabicyclo[2.1.1]hexanes on the physicochemical properties of bioactive compounds. For the comparison, we also used the corresponding carbocyclic core, bicyclo[2.1.1]hexane.
In summary, in two (fluxapyroxad, boscalid) out of four bioactive compounds, replacement of the ortho-substituted phenyl ring with 2-oxabicyclo[2.1.1]hexane led to a dramatic increase in water solubility by about one order of a magnitude.

Lipophilicity
To estimate the influence of the replacement of the ortho-substituted phenyl ring with saturated bioisosteres on lipophilicity, we used two parameters: calculated lipophilicity, log P, where P is the partition coefficient (clogP), and experimental lipophilicity, log D, where D is the distribution coefficient (logD).
Replacement of the phenyl ring with bicyclo[2.1.1]hexane either led to an increase of clogP (fluxapyroxad, boscalid; Fig. 3a,b) or did not affect it (phthalylsulfathiazole, lomitapide; Fig. 3c,d). However, in all four bioactive compounds, incorporation of 2-oxabicyclo[2.1.1] hexane instead of the ortho-substituted phenyl ring led to a decrease of the clogP index by about one unit.
In summary, in all tested compounds, replacement of the ortho-substituted phenyl ring with 2-oxabicyclo[2.1.1]hexane decreased the lipophilicity as measured by both clogP and logD indexes by 0.5-1.4 units.

Metabolic stability
The effect of saturated bioisosteres on the metabolic stability of bioactive compounds was complex and depended on the chemical  Fig. 3b). All three compounds, phthalylsulfathiazole and its two saturated analogues 32 and 33, were metabolically stable (Fig. 3c) Fig. 3d).
In summary, replacement of the ortho-substituted phenyl ring with 2-oxabicyclo[2.1.1]hexane in bioactive compounds improved metabolic stability (CI int ) in boscalid and fluxapyroxad; slightly decreased it in lomitapide; and did not affect it in phthalylsulfathiazole.

Bioactivity
Finally, we wanted to answer a key question: can 2-oxabicyclo[2.1.1] hexanes indeed mimic the ortho-substituted phenyl ring in real-world bioactive compounds? Fluxapyroxad and boscalid are marketed fungicides, developed by BASF, that have been approved for use in the United States and the European Union. Therefore, we measured their antifungal activity and compared it to that of their saturated analogues 28-31. In strict contrast to medicinal chemistry, the use of racemic mixtures in agrochemistry is common 50 ; therefore for the primary validation of the proof-of-concept, we directly studied the biological activity of the available racemic compounds 28-31 (Fig. 4).
First, we measured the antifungal activity of all compounds using the agar well diffusion method ( Supplementary Information, pages  238-243). Fluxapyroxad, and its saturated analogues 28 and 29, showed a similar trend in activity at the inhibition of fungi growth (Fig. 4a,b). The 2-oxabicyclo[2.1.1]hexane analogue 29 was active, but less potent compared to the original fungicide. Compound 29 and fluxapyroxad almost identically inhibited the growth of Fusarium oxysporum at high concentrations; however, at low concentrations, analogue 29 showed lower activity (Fig. 4a). Similarly, 29 and fluxapyroxad effectively inhibited the growth of Fusarium verticillioides at high concentrations; however, at low concentrations, only fluxapyroxad remained active, while analogue 29 did not (Fig. 4b).
Boscalid and both saturated analogues 30 and 31 also effectively inhibited the fungi growth (Fig. 4c,d). However, 2-oxabicyclo[2.1.1] hexane 31 was slightly less potent than boscalid at the inhibition of

1, 5-24
In situ generated * All reactions were performed on a gram scale. Diastereomeric ratio of esters 'a' in crude reaction mixtures after the photochemical step is given. Isolated yields of acids 'b' are given in three steps from allylic alcohols. a Product was isolated by crystallization from hexane-MeOtBu mixture. b Product was isolated by crystallization from acetone-water mixture. c Product was isolated by column chromatography.

Conclusion
The ortho-substituted phenyl ring (as well as meta and para isomers) is a basic structural element in chemistry. In this work, we synthesized, characterized and studied 2-oxabicyclo[2.1.1]hexanes as saturated bioisosteres of the ortho-substituted phenyl ring (Fig. 1c). These scaffolds were synthesized from readily available starting materials on a multigram scale. Crystallographic analysis revealed that these structures and the ortho-substituted phenyl ring indeed have similar geometric properties. Replacement of the ortho-substituted phenyl ring in bioactive compounds with 2-oxabicyclo[2.1.1]hexanes, in two out of four cases, dramatically improved water solubility (up to more than ten times) and metabolic stability. Moreover, in all four cases, such replacement also reduced lipophilicity by 0.5-1.4 clogP or logD units (Fig. 3b). In addition, the 2-oxabicyclo[2.1.1]hexanes 29 and 31 showed a similar antifungal activity compared to that of the original fungicides fluxapyroxad and boscalid.
Given the commonplace nature of the ortho-substituted phenyl ring in chemistry, we believe that its saturated bioisosteres described in this work will soon become very popular. One must always keep in mind, however, that the replacement of the phenyl ring in bioactive compounds with saturated isosteres can fail, if the phenyl ring is involved in integrations with the receptor: π-π stacking, π-amide   Article https://doi.org/10.1038/s41557-023-01222-0 stacking, π-Asp/Glu/Arg stacking, π to amide N-H, π to O-H, π to S-H, π to ammonium salts and so on. Therefore, the replacement of the phenyl ring in bioactive compounds with saturated isosteres must be careful and balanced 69 .

Online content
Any methods, additional references, Nature Portfolio reporting summaries, source data, extended data, supplementary information, acknowledgements, peer review information; details of author contributions and competing interests; and statements of data and code availability are available at https://doi.org/10.1038/s41557-023-01222-0.

General procedure for the photochemical [2 + 2] cycloaddition
The solution of diene 1 (16.79 g, 0.077 mol, 1.0 equiv.) and benzophenone (1.40 g, 0.0077 mol, 0.10 equiv.) in 850 ml of dry CH 3 CN (concentration = 0.091 M) was put into a standard chemical 1 l glass flask. The reaction mixture was degassed by the bubbling of argon for 15 min. The flask was closed by a septum and irradiated with luminescent UV lamps at 368 nm (24 lamps; Sylvania 368 Blacklight F25/T8/18/BL3368; each lamp has nominal power of 25 W; total power is 600 W), under stirring at room temperature for 48 h. The reaction mixture was concentrated under reduced pressure to provide the crude product 1a that was used in the next step (saponification) without any purification. NMR spectra were analysed with MestreNova (11.0.3-18688).

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

Data availability
Experimental data as well as characterization data for all new compounds prepared during these studies are provided in the Supplementary Information of this manuscript. The X-ray crystallographic coordinates for compounds 5b and 9b have been deposited at the Cambridge Crystallographic Data Center (CCDC) with accession codes 2166325 (5b) and 2166326 (9b). These data can be obtained free of charge from the Cambridge Crystallographic Data Center via www. ccdc.cam.ac.uk/structures/.