Recent Advances in Enantioselective Desymmetrizations of Prochiral Oxetanes

Abstract Strain relief of oxetanes offers a plethora of opportunities for the synthesis of chiral alcohols and ethers. In this context, enantioselective desymmetrization has been identified as a powerful tool to construct molecular complexity and this has led to the development of elegant strategies on the basis of transition metal, Lewis acid, and Brønsted acid catalysis. This review highlights recent examples that harness the inherent reactivity of prochiral oxetanes and offers an outlook on the immense possibilities for synthetic application.


Introduction
The strain embedded in small cyclic ethers has led to their wide applicationi nt he synthesis of chirala lcohols. [1] An elegant way to introduce asymmetry is through the synthesis of a prochiral cyclic ether followed by as ubsequentd esymmetrization step. In contrast to oxiranes (epoxides), [2] desymmetrization of oxetanes can occur at ad istant, sterically accessible position within the molecule, which offers au nique opportunity for the synthesis of chiral alcohols, especially those bearing bquaternary stereocenters (Scheme 1).
Since the late 1980s numerous strategies for desymmetrization have been developed utilizing transition metal catalysis, organocatalysis, and enzymatic methods. [3] Caused by their growingi mportance as bioisosters in medicinal chemistry,a plethora of methods for the synthesis of oxetanes came to fruition fueling the development of new desymmetrization reactions in recent years. [4] Duet ot heir high ring-strain, [5] oxetanes are prone to similar opening and atom-insertion reactions as oxiranes (Scheme 2, top). [6] Interestingly,t he oxygen atom in the oxetane ring is surprisingly Lewis basic, which can be explained by two contradictorye ffects. Whereas an arrow C-O-C angle leaves the oxygen lone pair less sterically hindered, it also changes the hybridizationo ft he ether's non-bondingo rbitals to highers -character. [7] It has been shown experimentally through heato fm ixing energies that oxetane comprises af avorable balance of the two effects (Scheme 2, bottom). [8] This manifestsi nasuperior electron donor ability compared to oxirane as well as oxolane (tetrahydrofuran) ando xane (tetrahydropyran). Similar trends have been observed for the binding to phenol, [9] iodine, [10] boront rifluoride, [11] and the protonation with nitric acid. [12] Accordingly,o xetanes are privilegeds tructures for activation with Lewis acids or Brønsted acids, and are prone to undergo subsequentr ing-opening reactions with an appropriate nucleophile. [1f, 13] In addition, insertion in the CÀO bondsi sp ossible, albeit so far not in an asymmetricf ashion. [14] Besides the common reactivity, the distorted CÀC s-bonds of the oxetane ring [15] offer opportunities to be activated by bcarbon elimination, rearrangement, and bond-insertion (Scheme 3). In these synthetic sequences, the two ether CÀO s bonds stay untouched and the reactivity of the oxetanes often resembles the one of the correspondingc yclobutanes. [16] However,r eaction development is affected by two major differences arising from the ether oxygen through inductive effects and competitive binding (due to the high Lewis basicity). Aside from that, no stereocenter can be created at the oxygen atom and current methods therefore focus on as tepwise desymmetrization approach via pinacol-type rearrangementa nd b-carbon elimination/addition (Scheme 3, left). The asymmetric direct insertion has so far not been achieved (Scheme3,r ight). The products from these "remote" oxetanea ctivationsa re highly diverse and synthetically useful,a nd clearlyd ifferentiate desymmetrizations of oxetanes from the related oxiranes.
In this review,c urrent methods that achieve enantioselective desymmetrizations are presented based on the strategies discussed above. First, the focus will be laid on the breaking of the CÀOb ond by nucleophilic ring-opening reactions arranged accordingt ot he type of nucleophile that is used. In the second section, the breaking of the remote CÀCb ond will be of interest and strategies involving transition metalsw ill be discussed before af inal conclusion and outlook will be drawn.

Nucleophilic Ring-OpeningR eactions
Oxetane ring-opening reactions by oxygen nucleophiles are important for heterocyclics yntheses and were employed in numerous studies. [17] Recently,C arreira et al. described an oxetane desymmetrization via the activationo ft he oxetane ring by addition of indium and boron-based Lewis acids. [18] Other types of Lewis acids have also been explored,s uch as Co III in an asymmetrict ransformation reported by the Jacobsen group in 2009. Here, ar ing-openingr eaction of oxetanol 1 to oxolane 2 catalyzed by Co III ·salen 3 was described (Scheme 4). [19] Ac ooperative bimetallic effect was found based on the superior reactivity of oligomeric Co III -catalyst 4 compared to the monomeric counterpart 3.I nterestingly,t his method also allowed the formation of quaternary stereocenters when 3,3-disubstituted oxetanes wereu sed. The driving force for this transetherification can be found in the differencei ns train energies between the oxolane and oxetaner ings (~20 kcal mol À1 ,c f. Scheme 2).
Based on initial findings on Brønsted-acid catalyzed ringopeningr eactions to lactones, [20] the Sun group elaborated on oxygen-based ring-opening reactions utilizing chiral phosphoric acids (CPAs) for enantioinduction (Scheme 5,top). [21] In their first communication, they reported oxetane ring-opening of 5 by 1,1'-spirobiindane-7,7'-diol (SPINOL) derived CPA 6 to give dioxane 7 in 99:1 enantiomeric ratio (er). Interestingly,t he authors were also able to access sterically encumberedd ioxanes such as 8 (Scheme 5, bottom).E ven when the size of the nucleophile tether wase xtended, the yields and observed enantioselectivitiess tayed excellent. Later,t he same group expanded the scope towardst he enantioselective synthesis of 1,4-Alexander Sandvoßs tudied chemistry at Westfälische-WilhelmsU niversitätM ünster as Studienstiftung scholara nd at Cardiff University (Prof. T. Wirth) as Erasmusf ellow. After a researchs tay in the groupo fP rof. A. Studer on on-surface reactions, he graduated in 2019 and joined the Wiest laboratory shortly after. Currently, he is working on strain-mediated reaction developmenti nvolving cyclic ethers and on applications in their synthesis.
Johannes Wiest attended the Universität Basel, Switzerland, where he obtained his M.Sc. in 2012. After ar esearch stay at the ScrippsR esearch Institute (Prof. D. G. Blackmond), he moved to Munich to pursue his Ph.D. at the Technische UniversitätM ünchen with Prof. T. Bach. He held aD FG research fellowship at Indiana University (Prof. M. K. Brown)b efore starting his independent career at the Westfälische-WilhelmsU niversitätM ünster.S ince 2020, he is an Assistant Professor at the Johannes Gutenberg-Universitäti n Mainz. Scheme4.Jacobsen's Co III -catalyzed intramolecularring-opening reactiontowards oxolanes. Tf = trifluoromethanesulfonyl. benzodioxepines such as 9 and 10. [22] Mechanistically,C PA 6 was suspected to increase the electrophilicity of oxetane 5 throughh ydrogen-bonding. The pseudo-C 2 -symmetric nature of catalyst 6 is enabling the high enantioselectivityt houghe ffective steric shielding by the large aryl units (see Figure1 for aG oodman and 3D representation).
The initial postulation that the acid motif of the CPAh as a dual role by activating both the oxetane and the nucleophile had to be revised due to recent theoretical calculations on the transition state geometry of this intramolecular process. It appears more likely that hydrogen-bonding is activating the oxetane ring, and an oncovalenti nteraction between the psystem of the CPAa nd the nucleophilic OH-group is involved. [23] This argumenta lso explainst he privileged role of large p-systems as flanking groups (e.g. 9-anthryl and 1-pyrenyl) in these types of catalysts.
As in prior examples, this transformation allows the control of stereocentersr emote to the reaction site. The observed enantioselectivity was explained by the transition state geome-try,w hich differs depending on the ability of the substituents to undergoh ydrogen bondingt ot he catalyst, and the respective arrangement of its steric backbone (cf. Figure 1). Based on the aforementioned results, the Sun and Houk groupsc ollaborated to study the mechanism of intramolecular ring-opening reactionsw ith protected oxetanol 16 as an exemplary substrate (Scheme 7). [26] The influence of different CPA'sw as investigated in this study,b ut these acids wereg enerally not effective in catalyzing the transformation, presumably due to their insufficient acidity.H ence, chiral phosphoramidate 17 havinga highera cidity was identified. With this catalyst, the reaction proceeded in high yield and in an enantioselective fashion. The proposed mechanism entailing ani ntramolecular nucleophilic attack of the protected sulfur (intermediate 19)a nd a subsequenti ntramolecular protecting group exchange (via 20) is based on density functional theory (DFT) calculations and cross-over experiments.M oreover,t his reaction was tolerant to different substituents at the 3-position of the oxetaner ing and even allowed thiane 21 to be accessedw hen adjusting the protecting group.
Ring-openingr eactions with carbon nucleophiless uch as alkyl lithiums or enolatesa re rare, even in ar acemic fashion. [27] In 1996, the first enantioselective desymmetrization using a lithiated carbon nucleophile was reported by Tomioka et al. In this study, 3 -phenyloxetane (11)w as treated with stoichiometric amountso fb oron trifluoride along with chiral ligand 22 (Scheme 8, top). [28] It was postulated that polyether 22 coordinates the lithium ion in ar igid bicyclic structure (such as 25), which directs the attack of the phenyla nion and thus acts as the origin of enantioselectivity for the formation of chiral alcohol 23 in 73.5:26.5 er.I nterestingly,t he enantiomeric excess was also influenced by the type of Lewis acid, with boron trifluorideb eing the best choice activating the oxygen atom of the cyclic ether beforet he nucleophilic attack. Different nucleophiles such as lithium phenylacetylide were viable and provided the correspondingp roduct 24 in high yield, albeit moderate selectivity.T his example showst he diversity of products that can be generated, but also highlights the limits of current protocols in terms of enantioselectivity.P romising results were observed in av ery recent study by Sun et al. concerning soft Scheme6.Sun's intermolecular ring-opening reaction using sulfur nucleophiles.
Regarding nitrogen nucleophiles, ring-openingr eactions of oxetanesa re challenging indicatedb yt he harsh conditions used in early studies by Ziemann and Gregory. [30] More recent reports are typically based on an intramolecular approacht argeting heterocyclic rings. In this regard,t he Kuduk and Stew-ard groups reported one-pot procedures that attach 3-aminooxetane to the aromatic scaffold followed by as ubsequent ring-opening reaction to forge N-heterocycles such as quinazolines or indazoles in aracemic fashion. [31] In terms of enantioselectivity,t he Sun group reportedt he synthesis of tetrahydroisoquinolines through desymmetrization of 2-oxetanylbenzaldehyde 29 in 2013 (Scheme 9, top). [32] This one-pot procedure is high yieldinga nd highly enantioselective, although the reaction is limited to electron rich aryl amines.T oc arry out this formal reduction,t he addition of Hantzsch ester 31 was required. The authors suggest two possible reactionp athways (Scheme 9, bottom left). In path a, the reactionp roceeds via  Figure 1). [21] To showcase the synthetic applicability,t he Zhu and Sun groups applied this strategy in am ulti-component aza-Diels-Alder reaction with indoles to form highly complex polycyclic indolines such as 38 in good diastereo-and enantioselectivity (Scheme 9, bottom right). [33] Besides the common nucleophiles oxygen,s ulfur,c arbon, and nitrogen, desymmetrization of oxetanes has been achieved with phosphorus-species in context of ligand design. [34] Despite the importance of chiral phosphines as ligands, no enantioselective variant hasb een reported so far.F urthermore, halogensa re known to undergo nucleophilic attack at oxetane rings. [35] An enantioselective variant of this ring-opening reac-tion was reportedb yS un et al. using CPA 42 for enantioinduction (Scheme 10, top). [36] To this end, aw ide range of substituents on the oxetane ring were employed (e.g. phenyl, 11 to 39 and benzylether, 40)a sw ella s3 ,3-disubstitution, which gave rise to fully substituted stereocenters in as tereocontrolled fashion (e.g. 41). Mechanistic experiments omitting CPA 42 showede ffective background reaction, which made optimization of the reaction parameters particularlyc hallenging. The key to success was the small and continuous releaseo fw ater from wet molecular sieves, which triggered the continuous release of HCl (via the reaction of water with trimethoxy silyl chloride) and allowed the ring-opening to occur in ah ighly enantioselectivef ashion.T og et further insight into this reaction, theoreticalc alculations were performed confirming ab ifunctional activation mode of the CPAt hrough coordination of its Lewis acidic side to the oxetane ring and its Lewis basic side to the HCl( cf. Figure 1). [23a] An elegantr ing-opening with bromide as nucleophile was reported by Jacobsen and co-workersi n2 020 (Scheme 10, bottom). [37] Trimethylsilyl bromide acted as the bromide source,w hiche nhanced the electrophilicity of the oxetane through silylation of the Lewis basic oxygen and allowedt he formationo fb romo ether 43 from phenyloxetane 11.F unctional groups such as benzyle ther 44 were also compatible with this reaction protocol. Mechanistically,t he chiral squaramidic catalyst 46 was proposed to interact with the bromide via its hydrogen bondingmotif and with the substrate through its Lewis basic amine functionality (see structure 45). Thus, excellent enantioselectivity was achieved for the delivery of the bromide to the oxetane. Kinetic isotope effect (KIE) analysis in-dicatedt hat silylation of the oxetane ring was reversible and the bromide ring-opening/CÀOb ond cleavagee nantiodetermining.

Remote Oxetane Desymmetrization
Besides Lewis acids and Brønsted acids, desymmetrization reactions of oxetanes can be promoted by transition metal catalysts activating the remote CÀCb ond. In this context, Zhang et al. reported ar ing expansion reaction of 3,3-disubstituted oxetane 47 with ar hodium catalyst and alkyne 49 (Scheme 11). [38] The reactionl ikely proceeds via the coordination of the catalyst to the hydroxy and the aryl moiety of 47 (intermediate 51). After subsequent b-carbon elimination to intermediate 52,m igratory insertion across the alkyne followed by ring-closure and protoderhodation( from 54)g ives rise to cyclic ether 50.T he geometry of intermediate 53 determines the stereoisomeric outcomeo ft he reaction, which was induced by BINAPINE ligand 48.O ther symmetric alkynes were also viable in this synthetic sequence (e.g. 55). Ar elated,r acemic example of ar ing-expansion reaction wasr eported by Miura et al.,i nw hich aC ÀH-activation strategy was combined with the ring-opening and closing reactionoft he oxetane. [39] Desymmetrization reactions based on 2,2-disubstitution at the oxetane ring are less studied and there is only one report by the group of Njardarson to date. [40] In this study,2 -monosubstituted oxetanes were the primary objecto fi nvestigation. However,t oexplore the mechanism the authors subjected symmetric divinylo xetane 56 to their optimal reaction condi-tions (Scheme 12). It was suspected that the reaction proceeds through as tepwise mechanism,w here the copper is activating the oxetane as aL ewis acid, followed by ring-opening. The cation 59 is delocalized through the allylic system and by anucleophilic attack of the oxygen atom, the oxolane ring is formed. Higher activity of copper-catalyst 57 comparedt ot he phosphoramidate 58 was observed with ac ontradictory influence on the enantioselectivity of the product 60.
Remote desymmetrization of oxetanes can also be achieved through rearrangement.I nt his context, semipinacol rearrangements are particularly popular,b ut are typicallyp erformed on cyclobutanol ring expansions. [41] Regarding oxetanes, the Youg roup recently investigated the behavior of oxetaned erivatives in such ar earrangementr eaction (Scheme 13). [42] Therefore, allylic alcohol 61 was submitted to ar eaction sequence based on electrophilic chlorine source 62.T he reaction likelyp roceeds via the enantioselective formation of chloronium 66 from the addition of an electrophilic chlorine-atomt ot he CÀCd ouble bond of the allylic alcohol 61.C hloronium 66 can then undergo as emipinacolr earrangement generating cyclic ketone 63 and 64,r espectively.T he addition of N-Boc-(L)-phenylglycine( NBLP)a nd the phthalazine adduct of dihydroquinine ((DHQD) 2 PHAL) was found to be crucial for the enantioinduction in this reaction (presumably via coordination of Cl + before alkene attack, see 65).

Conclusion
This review highlightst he synthetic utility of prochiral oxetanes, which is based on their inherent ring-strain and Lewisbasicity.E legant desymmetrization strategies have been established for 3-substituted,3 ,3 as wella s2 ,2-disubstituted oxetanes, which give rise to ad iversity of products.T he enantioselectivity of theser eactions can be controlled through sophisticated choice of Lewis acids or Brønsted acids, with CPAs being the most widely used catalystt od ate. Some of the presented reactionsa llow control over remote stereocenters, which marks au nique strategy in the synthesis of chiral alcohols and ethers.H owever,w hen compared to the synthetic impact of oxiranes (epoxides), oxetanes are still far less studied and their application in synthetic endeavors lags behind. Additionally, the possibility of oxetanes to undergo other reactions than nucleophilic ring-openings at the a positionw ere only recently explored indicated by three examples from the Zhang, You, and Njardarson groups. Afurther reason for the lackofapplication is the easy,b ut stepwise formation of the oxetane ring that cannotc ompetew ith epoxidations. The comparable physical properties of oxetane in terms of ring strain as well as its significance as ab ioisoster are however promising and set the basis for future explorations to fully exhaust the immense possibilities of oxetanes for synthesis and pharmaceuticala pplications. In particular desymmetrization reactions resemble an attractives trategy,a st hey allow oxetanes to be converted to highly complex scaffolds bearing quaternary (all-carbon) stereocenters in astep-economic fashion.