O O O

O O Oxetane Drug Development, Synthesis & Applications!

John Thompson Dong Group Literature Seminar September 24th, 2014 OH O O

O Overview!

1. Drug Discovery & Pharmaceutical Interest

2. Chemical Properties & Synthesis of Oxetane Rings

3. Applications of Oxetane Rings

i. Ring Opening for Complex Molecule Synthesis

ii. Organometallic Chemistry O R'O OH R

O NH O O O OH O OH O O O

R = Ph, R' = Ac, Paclitaxel (Taxol) R = OtBu, R' = H, Taxotere (Docetaxel)

All marketed drugs containing the oxetane ring come from the Taxane family of natural products

Computational studies show the oxetane moiety providing: (1) rigidification of the overall structure (2) H-Bond acceptor for a threonine-OH group in binding pocket

The full extent of the oxetane biological role is still unclear Other Oxetanes in Natural Products

NH2 N Me O N O HO O H N N CO2H Me O O C H O 5 11 OH O OH O O O Me O OH Me Me OH OH Thromboxane A 2 Merrilactone A Maoyecrystal I Oxetanocin A promotes vasoconstriction/ stimulates rat neuron growth anti-cancer HIV Inhibitor platelet aggregation

Me Me NH Me Me NH 2 H N NH O O 2 2 O CO2H Me O O OH Me O O MeO2C Me O Me Bradyoxetin Oxetin Dictyoxetane gene regulation in soybean Mitrephorone A herbicidal / antibacterial polycyclic diterpenoid high cytotoxicity

Angew. Chem. Int. Ed. 2010, 49, 9052.! Medicinal Chemistry

§ Compound property optimization is a major hurdle for drug discovery § Small molecules that can be easily added onto and change compound properties in predictable ways are highly valued § The oxetane ring is a very small molecule whose properties in the past decade have shown far reaching advantages for biological modulation

– This compound has been neglected due to difficult synthetic access and concerns about chemical and metabolic stability O

Key Figures Erick M. Carreira Klaus Müller ETH Zürich Angew. Chem. Int. Ed. 2010, 49, 9052.! Angewandte Oxetanes Chemie and safety screens to examine, refine, as well as predict chemical space beyond the structures that initiated the druglike attributes of the candidate structures. As predictive research endeavor. The collection of molecules generated tools and assays become available the process is conducted result in a multidimensional network of structure–activity increasingly earlier in the drug discovery process so as to relationships that provide the basis for identifying and Oxetan-3-one: Chemistry and Synthesis 221 minimize attrition at the later (and more expensive) preclin- defining the pharmacophore. ical and clinical phases as a result of toxicity or lack of Our own studies wereaffords initiated regioselectively by the the idea corresponding that oxetanes oxetanes in good yield.51 An important difference efficacy. In medicinal chemistry it is common to identify could be viewed as a gembetween-dimethyl the various equivalent, approaches is wherein that in the the Williamson approach, stereochemical issues are addressed separately from the ring-closing event. By contrast, in the processes that commence with robust surrogates for typical functional groups. These surro- two methyl groups arecarbonyl bridged substrates, by an control oxygen must be atom. exercised It over was the generation of stereocenters during the ring gates are modules that may share a limited set of key reasoned that the polarclosure oxygen step. Although bridge the would approach beinvolving able intramolecular to Williamson ether synthesis permits access to a range of substituted oxetanes, the latter two methods necessarily lead to oxetanes that properties with the parent functional group, such as acidity/ compensate for the intrinsicincorporate lipophilicity substitution at C-2. of the methylene basicity, size, and conformational biases. For example, com- groups, such that the net change in lipophilicity would be mon replacements for carboxylic acids (RCOOxetanes2H) and gem- small to or Replace even vanish (Figure13.4 Common 1). OXETANES Further IN research DRUG Functionalities DISCOVERY on oxetanes dimethyl groups are tetrazoles (RCN4H) and cyclopropanes, has undergone additionalThe twists calculated and van turns der Waals that volumes have of revealed trimethylene oxide and propane are essentially the same. respectively. Although on one level these surrogates are a number of other intriguingConsistent structural with this feature, relationships. the measured Thus, partial formolar volumes in water are similar.52,53 This leads perceived as interchangeable in the discovery process, it is to the hypothesis that it may be a fruitful exercise to consider the replacement of gem-dimethyl § Replacement of gem-dimethyl groupsgroups (t-butyl with oxetanes = methyl (Figure substituted 13.5). This would gem-dimethyl lead to geometrically group) similar structures with well appreciated that the equivalence is merely superficial, markedly different pharmacokinetic properties. In a related fashion, the tert-butyl group can be because each functional group has its own intrinsic structural considered a gem-dimethyl substituted ethyl group. Thus, in a model construct wherein oxetanes would function as gem-dimethyl group equivalents, 3-methyl 3-oxetanyl linked units would serve or electronic characteristics. Nonetheless, the replacement as tert-butyl surrogates. Both gem-dimethyl as well as tert-butyl groups can be found in many groups permit the medicinal chemist to navigate among pharmaceutical compounds. A search of the World Drug Index reveals 714 entries for molecules various structural classes and bridge to diverse regions of incorporating a tert-butyl group, and 69 such compounds are currently in the market. Tert-butyl groups are often used to fill hydrophobic pockets of a certain target; gem-dimethyls are often – Why are these groups so commonintroduced toin protect drugs? metabolically Steric labile hindrance positions. Inprevents both cases, chemicalhowever, their introduction is or metabolic liabilities of nearbyassociated functional with decreased groups water solubility and elevated lipophilicity. Higher lipophilicity in general Johannes Burkhard was born in Zurich, makes the molecule as a whole a better substrate for oxidative metabolic enzymes. However, the Switzerland in 1983. In 2006, he received • More than 10% of all launchedoxetane groupdrugs is considerablycontain at reduced least in one lipophilicity. gem-dimethyl In a series ofgroup studies, we were able to his Masters in chemistry from the ETH demonstrate that introduction of the oxetane into a model compound can lead to increased water solubility in excess of 4000-fold over the parent compound. Moreover, the 3,3-disubstituted oxet- Zurich under the supervision of Franç–ois However, replacing H for Me increasesanes were shown lipophilicity to be stable in (maya pH range have from adverse 1 to 10 and toeffects) enhance metabolic stability of the Diederich. For his PhD studies he stayed at parent molecules significantly in standard liver microsomal assays.19 the ETH, where he joined the research In order to make use of these potential beneficial properties of oxetanes as surrogates of gem- group of Erick M. Carreira in 2007.§ Oxetane His & gem-dimethyl groups havedimethyl near and tert equivalent-butyl groups, partialsynthetic methodsmolar hadvolumes to be developed in water that allow access to a variety of 3,3-disubstituted oxetanes. In principle, there are at least two approaches that could be pursued research is focused on the development– ofLeads to geometrically similar structures with markedly different pharmacokinetic small heterocyclic building blocks and theirproperties Figure 1. Oxetane analogies in medicinal research. R1 evaluation in the context of drug discovery. Me R1 Me Me R1 – World Drug Index (2008) Me Me OH Me • 714 molecules with t-butyl groups H Me • 69 in market Georg Wuitschik, born in Bad Tölz, Ger- Klaus Müller, born in 1944 near Lucerne, R1 R1 many in 1980, received his Diploma in Switzerland, studied chemistryR1 at ETHO Zur- O O chemistry in 2004 from the TU Munich for ich (PhD with Albert Eschenmoser). AfterOH Me Angew. Chem. Int. Ed. 2010, 49, 9052.! H work carried out in the group of Barry M. several years in the US (Chicago, Harvard), Trost under the supervision of Wolfgang A. he returned to ETH Zurich, where he R3 R2 R Herrmann. He then joined the group of completed his habilitationxy in Physical and R3 xy2 Me Me Erick M. Carreira, and received his PhD in Theoretical Organic Chemistry in 1977. In O 2008. He is currently working as an 1982 he joined F. Hoffmann–La Roche, FIGURE 13.5 Relevant structural motifs for the replacement of a gem-dimethyl group by an oxetanyl unit. Alexander-von-Humboldt postdoctoral fellow Basel, to set up programs in molecular in the group of Steve V. Ley, where he works modeling, structural biology, and on natural product synthesis. bioinformatics. In 1990 he became extra- ordinary professor at the University of Basel. Since his retirement in Spring 2009, he has continued as a chemistry consultant for Roche and holds a teaching position at ETH Zurich.

Mark Rogers-Evans was born in London in Erick M. Carreira was born in Havana, 1963. After his PhD and postdoctoral Cuba, in 1963. He received his BSc from studies in England (Brian Marples and the University of Urbana-Champaign work- Raymond Bonnett) and Canada (Victor ing with Scott Denmark, and his PhD from Snieckus), he joined Hoffmann—La Roche Harvard University working under the direc- in Switzerland (1996) as a Process Research tion of David A. Evans. After postdoctoral Chemist, transferring to Discovery Chemistry research at the California Institute of Tech- in 2001. In 2009 he joined the newly nology with Peter Dervan, he joined the formed Roche Chemistry Technologies & faculty there. Since 1998, he has been full Innovation group and has co-authored over professor at ETH Zürich. 60 patents and publications.

Angew. Chem. Int. Ed. 2010, 49, 9052 – 9067  2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.angewandte.org 9053 Angewandte Oxetanes Chemie

bene precursors were necessary to obtain high trans/cis ratios. oxygen atom in oxetanes and on the carbonyl oxygen atom The best results were observed for the 1,1-dicyclohexyl- display comparable spatial arrangements, and both function- substituted ethyl diazoacetate (102). The diastereomeric ratio alities are polarized similarly. Consequently, an analogy can Oxetanesof the products to was Replace largely determined byCommon the character of the Functionalitiesbe readily drawn (Figure 7, left). According to this logic, copper complex. Application of the (R,R)-bis(azaferrocene) replacement of a carbonyl group with an oxetane ring could ligand 104 led to the formation of tetrahydrofurans with trans- configured substituents, whereas the same reaction with the § Carbonyl Surrogate! other enantiomer of the ligand afforded predominantly cis- configured products. – Aldehydes (sterically accessible Michael acceptors) or acyl halidesAngewandte are precluded from drug discovery Oxetanes 5. Property ChangesChemi throughe Oxetanes As outlined in the introduction, oxetanes can play a bene precursors were necessary to obtain high trans/cis ratios. oxygen atom in oxetanes and on the carbonylsignificant oxygen role atom in the optimization of key pharmacokinetic § More stable groups: esters, ketones, and amides have liabilitiesFigure 7. Comparisonrelated of oxetanes and other polar functionalities.[2b] The best results were observed for the 1,1-dicyclohexyl- display comparable spatial arrangements,properties and both function-of drug candidates. This section focuses on the substituted ethyl diazoacetate (102). The diastereomeric ratio alities are polarized similarly. Consequently,– incorporationmany an analogyenzymes ofcan oxetanes can hydrolyze onto frameworks commonly seen of the products was largely determined by the character of the be readily drawn (Figure 7, left). According in medicinal to this chemistry logic, and their effect on the underlying be beneficial, since carbonyl groups, such as in aldehydes, [4] copper complex. Application of the (R,R)-bis(azaferrocene) replacement of a carbonyl group with anscaffold. oxetane ring could ketones, or Michael acceptors, are generally absent in drug § Alpha-carbonyl compounds ease of deprotonation can destroydiscovery stereocenters because of their. inherent chemical and metabolic ligand 104 led to the formation of tetrahydrofurans with trans- liability. Moreover, the relative ease of a-deprotonation in configured substituents, whereas the same reaction with the 5.1. Analogy of Oxetanes to gem-Dimethyl Groups ketones and aldehydes renders stereogenic centers at this other enantiomer of the ligand afforded predominantly cis- position sensitive towards epimerization. Carboxylic acid configured products. The introduction• ofExchange steric hindrance spirocyclic often deflectsoxetane derivatives,for morpholine such as! esters and amides, are chemically rather chemical[69] or metabolic[70] liabilities from nearby functional stable but are prone to enzymatic cleavage in an organism. groups. Especially in the• case 17 ofmarketed metabolically drugs unstable with morpholineTherapeutic units agents with oxetanes in place of the carbonyl 5. Property Changes through Oxetanes methylene groups as in a• benzylic Majority position, all degrade it is common group might, therefore, provide intriguing alternatives. practice to block metabolic attack by the introduction of a As shown in Figure 7 (right), a spirocyclic oxetane could As outlined in the introduction, oxetanes can play a gem-dimethyl unit.[71] It is interesting that more than 10% of serve as a viable substitute for morpholine, another wide- significant role in the optimization of key pharmacokinetic all launched drugs contain at least one gem-dimethyl group, spread moiety in pharmaceutical chemistry. Morpholine is [2b] properties of drug candidates. This section focuses on the Figure 7. Comparison of oxetanes and other polarthus functionalities. highlighting its relevance in drug discovery.[4] However, often used as a hydrophilic solvation anchor for lipophilic incorporation of oxetanes onto frameworks commonly seen for a typical small molecule in medicinal chemistry the compounds,Angew. but Chem. often Int. Ed. appears 2010, 49 to, be9052. the! target of oxidative in medicinal chemistry and their effect on the underlying be beneficial, since carbonyl groups, suchreplacement as in aldehydes, of hydrogen atoms by methyl groups leads to a metabolic attack. scaffold. ketones, or Michael acceptors,[4] are generallysignificant absent increase in drug in lipophilicity, which in turn may adversely affect its physicochemical and pharmacological discovery because of their inherent chemical and metabolic properties.[72] Moreover, the gem-dimethyl group can itself 5.3. Influence on the Basicity of Proximal Amines liability. Moreover, the relative ease of a-deprotonation in become a target of metabolic degradation.[73] The case 5.1. Analogy of Oxetanes to gem-Dimethyl Groups ketones and aldehydes renders stereogenicdescribed centers for at the thisgem-dimethyl group exemplifies the situa- While the oxygen atom of an oxetane can donate electron position sensitive towards epimerization.tion Carboxylic with various acid other substituents, which, similar to the gem- density as a Lewis base, the oxetane motif itself acts as an The introduction of steric hindrance often deflects derivatives, such as esters and amides, aredimethyl chemically group, rather intrinsically possess high lipophilicity, which electron-withdrawing group on neighboring functional [69] [70] chemical or metabolic liabilities from nearby functional stable but are prone to enzymatic cleavageis generally in an organism. undesirable in drug discovery.[72] Therefore, a groups. This distance-dependent inductive effect of the groups. Especially in the case of metabolically unstable Therapeutic agents with oxetanes in placestable, of the compact carbonyl module with reduced lipophilicity and oxetane can be used to temper the basicity of a proximal methylene groups as in a benzylic position, it is common group might, therefore, provide intriguingsusceptibility alternatives. to metabolic attack is desirable. By bridging amine. Data that has been accumulated from oxetanes 58, 60, practice to block metabolic attack by the introduction of a As shown in Figure 7 (right), a spirocyclicthe two oxetane methyl could groups of the gem-dimethyl unit with an 106, and 107 as well as the control compound 105 show gem-dimethyl unit.[71] It is interesting that more than 10% of serve as a viable substitute for morpholine,oxygen another atom, the wide- so-formed oxetane fulfills precisely these distinct patterns in amine basicities (Figure 8).[2a] The basicity all launched drugs contain at least one gem-dimethyl group, spread moiety in pharmaceutical chemistry.requirements. Morpholine The four-membered is ring occupies a similar, or reduction may be close to 3 pKa units if the oxetane is grafted thus highlighting its relevance in drug discovery.[4] However, often used as a hydrophilic solvation anchoreven smaller, for lipophilic volume than a gem-dimethyl group, as evident in the a position to the amine. It is continuously attenuated for a typical small molecule in medicinal chemistry the compounds, but often appears to be theby target comparison of oxidative of the partial molar volumes of oxetane with increasing topological distance to the oxetane. Even at 3 1 3 1 replacement of hydrogen atoms by methyl groups leads to a metabolic attack. (61.4 cm molÀ ) and propane (70.7 cm molÀ ) in water. The large distances, for example, in the d position to the oxetane, slightly smaller volume observed for oxetane is probably as in structure 107, a decrease of 0.3 pK units is observed. significant increase in lipophilicity, which in turn may a attributable to its hydrogen-bonding capacity.[74] adversely affect its physicochemical and pharmacological [72] gem- properties. Moreover, the dimethyl group can itself 5.3. Influence on the Basicity of Proximal Amines 5.4. Lipophilicity and Aqueous Solubility [73] become a target of metabolic degradation. The case 5.2. Oxetanes—Bulky and Polar described for the gem-dimethyl group exemplifies the situa- While the oxygen atom of an oxetane can donate electron Having seen that the presence of an oxetane can markedly tion with various other substituents, which, similar to the gem- density as a Lewis base, the oxetane motifWhile itself actsoxetanes as an were initially explored because they could influence the basicity of a proximal amine, it was important to dimethyl group, intrinsically possess high lipophilicity, which electron-withdrawing group on neighboringprovide steric functional bulk without increasing the lipophilicity, other assess its influence on the lipophilicity of the underlying is generally undesirable in drug discovery.[72] Therefore, a groups. This distance-dependent inductiveopportunities effect of arose the when oxetanes were envisioned as scaffold. To aid comparisons with neutral compounds (for stable, compact module with reduced lipophilicity and oxetane can be used to temper the basicitycarbonyl of a surrogates. proximal The electron lone pairs on the ether example, lactam derivatives) and to avoid potential compli- susceptibility to metabolic attack is desirable. By bridging amine. Data that has been accumulated from oxetanes 58, 60, the two methyl groups of the gem-dimethyl unit with an 106, and 107 as well as the control compoundAngew. Chem.105 Int. Ed.show2010, 49, 9052 – 9067  2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.angewandte.org 9061 oxygen atom, the so-formed oxetane fulfills precisely these distinct patterns in amine basicities (Figure 8).[2a] The basicity requirements. The four-membered ring occupies a similar, or reduction may be close to 3 pKa units if the oxetane is grafted even smaller, volume than a gem-dimethyl group, as evident in the a position to the amine. It is continuously attenuated by comparison of the partial molar volumes of oxetane with increasing topological distance to the oxetane. Even at 3 1 3 1 (61.4 cm molÀ ) and propane (70.7 cm molÀ ) in water. The large distances, for example, in the d position to the oxetane, slightly smaller volume observed for oxetane is probably as in structure 107, a decrease of 0.3 pKa units is observed. attributable to its hydrogen-bonding capacity.[74]

5.4. Lipophilicity and Aqueous Solubility 5.2. Oxetanes—Bulky and Polar Having seen that the presence of an oxetane can markedly While oxetanes were initially explored because they could influence the basicity of a proximal amine, it was important to provide steric bulk without increasing the lipophilicity, other assess its influence on the lipophilicity of the underlying opportunities arose when oxetanes were envisioned as scaffold. To aid comparisons with neutral compounds (for carbonyl surrogates. The electron lone pairs on the ether example, lactam derivatives) and to avoid potential compli-

Angew. Chem. Int. Ed. 2010, 49, 9052 – 9067  2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.angewandte.org 9061 Oxetanes to Replace Morpholine

3228 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 8 Wuitschik et al.

Figure 2. Marketed drugs containing a morpholine unit susceptible to metabolic oxidation (A) and spirocyclic oxetanes that could be used in place of a morpholine unit (B). The spirocyclic oxetane amines represent interesting building blocks in view of the characteristic spatial arrangementReplacement of polar surface domainsof gem- (thedimethyls light-blue, barscarbonyls, mark linear and or reclined morpholine arrangements units of for local oxetane polar units) derivatives and as possible in analogues for corresponding ketoamines and lactams (C). several tests all show higher metabolic stability!

J. Med. Chem. 2010, 53, 3227.!

Figure 3. Averaged structural parameters obtained from nine X-ray structures of 3,3-disubstituted oxetane derivatives and two low- temperature X-ray structures of the parent oxetane, as extracted from the CSD.11 sterically accessible Michael acceptors,9 or acyl halides is and its structural impact in diverse topological environments. precluded in drug discovery. Moreover, more stable func- Collectively, this set of data provides valuable insight into tional groups such as esters, amides, or ketones can have nonbonded interactions involving oxetanes and conforma- intrinsic liabilities associated with them in the drug discovery tional effects induced by these small rings. process. Numerous enzymes can hydrolyze esters and amides The average structural parameters for a 3,3-disubstituted or reduce ketones. Furthermore, the relative ease of R-depro- oxetane are shown in Figure 3. The oxetane displays a slight tonation in carbonyl compounds can render stereogenic widening of the exocyclic C-C-C bond angle and a small centers at this position susceptible to epimerization. In such degree of puckering (measured by the dihedral angle θ). In settings an oxetane unit in place of the carbonyl group, as the gas phase, microwave spectroscopy data suggest an generally depicted in Figure 1 and illustrated specifically for effectively planar structure for the parent oxetane14 with a cyclic ketoamines or lactams in Figure 2, can provide an low pucker-inversion barrier. Substitution at the 3-position interesting structural alternative. leads to increased eclipsing interactions with the adjacent Aqueous solubility of lipophilic scaffolds is often improved methylene groups, and therefore, more pronounced pucker- by the attachment of a morpholine unit, even though morpho- ing of the oxetane ring is found in many of these structures. line is known to be a likely target of oxidative metabolism. Nevertheless, the puckering angle remains quite small, which There are 17 marketed drug substances that contain the is important with regard to the structural analogy between an morpholine moiety. For 13 of them metabolic data have been oxetane and a carbonyl unit. published,10 and in 8 compounds oxidative degradation of An oxetane unit exerts a distinct influence on the con- the morpholine ring (Figure 2)11 is the predominant meta- formational preferences of the backbone chain to which it is bolic pathway. The class of spirocyclic oxetanes shown in grafted (Figure 4). Whereas antiperiplanar conformations Figure 2 may be taken as structural analogues of morpholine are favored for an unsubstituted aliphatic chain (τ = 0), all in their ability to increase solubility while maintaining stability three staggered conformations are approximately equally toward oxidative metabolism. However, these oxetane deri- populated at the site of a gem-dimethyl group (τ =0°, 120°, vatives are intrinsically interesting building blocks through and -120°). For a 3,3-disubstituted oxetane, the synclinal their distinct spatial distribution of polarity. (gauche) conformations (τ = 120° and -120°) are clearly preferred over an antiperiplanar arrangement of the chain. This is likely a result of the small oxetane C-C-C valence Results and Discussion angle ( 84°) that leads to an increase of steric repulsion in Structure and Conformation of Oxetanes. The analogies the antiplanar∼ backbone arrangement compared to that of a regarding 3-substituted oxetanes depicted in Figure 1 rely on gem-dimethyl group.15 limited structural knowledge12 available from the Cambridge The analogy of an oxetane with a gem-dimethyl group Structural Database (CSD).13 Additional X-ray structures draws on the experimental fact that the partial molar volume were obtained in the course of this work highlighting oxetane of oxetane in water (61.4 cm3/mol at 25 °C)17 is comparable 3228 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 8 Wuitschik et al.

Minireviews Figure 2. Marketed drugs containing a morpholine unit susceptible to metabolic oxidationM. Rogers-Evans (A) and spirocyclic et oxetanesal. that could be used in Chemical Propertiesplace of a morpholine unit (B). The spirocyclic oxetane amines represent interesting building blocks in view of the characteristic spatial arrangement of polar surface domains (the light-blue bars mark linear or reclined arrangements of local polar units) and as possible analogues for corresponding ketoamines and lactams (C). corresponding carbonyl compounds are even more hydro- philic. Thus, the oxetane positions itself between a carbonyl § Structure and a gem-dimethyl group, but is somewhat closer to the Minireviews – Wider C-C-C bondM. Rogers-Evans et al.former. The observed trend in lipophilicity manifests itself in – Slight puckering intrinsic aqueous solubilities.[2a,b] A carbonyl compound, corresponding– carbonylGas phase compounds suggests are even planar more hydro-under the same experimental conditions, is usually more philic. Thus, the oxetane positions itself between a carbonylsoluble than its oxetane or gem-dimethyl analogues. The – 3-substitution increasesFigure 3. Averaged puckering structural parameters (due to obtained eclipsing from nine interactions X-ray structures of 3,3-disubstituted oxetane derivatives and two low- and a gem-dimethyl group, but istemperature somewhat X-ray closer structures to of the thechanges parent oxetane, seen in as the extracted series from58, the60 CSD., 10511,–107, and 116 emphasize former. – Strain energy is 107 kcal/mol sterically accessible Michaelthe acceptors, effects9 or the acyl introduction halides is of theand polar its structural ethereal impact oxygen in diverse atom topological environments. The observed trend in lipophilicityprecluded manifests in drug discovery.O itself incan Moreover,O have, especiallymore stable when func-O employedCollectively, away this from set of other data polarprovides valuable insight into [2a,b] O Figureintrinsic 8. Change aqueous in the pKa value solubilities. of an aminetional uponA carbonyl replacement groups such compound, of as a esters,functional amides, or ketones groups. can When have situatednonbonded close interactions to a basic involving amino oxetanes and conforma- methyleneunder group the with same an oxetane experimental ring (difference conditions,intrinsic in brackets). liabilities is usually associated more withgroup, them the in the observed drug discovery gain intional solubility effects induced may be by these moderate small rings. process. Numerous enzymes can hydrolyze esters and amides The average structural parameters for a 3,3-disubstituted pKa valuessoluble are amine than basicities its oxetane measured or gem spectrophotometrically-dimethylRing Strain analogues. at114 The107 23 5 or reduce ketones. Furthermore,because the relative of the ease reduction of R-depro- of basicityoxetane and are concomitant shown in Figure increase 3. The oxetane displays a slight 248C in H O. (kJ/mol) changes2 seen in the series 58, 60, 105tonation,–107, and in carbonyl116 emphasize compoundsin the can intrinsic render lipophilicity stereogenic (Figurewidening 10; 60 of theand exocyclic106). However, C-C-C bond angle and a small the effects the introduction of the polarcenters ethereal at this position oxygen susceptible atom to epimerization. In such degree of puckering (measured by the dihedral angle θ). In can have,§ especiallyAchiral when employed substitutedsettings away an on from oxetane 3-position other unit inpolar place! of the carbonyl group, as the gas phase, microwave spectroscopy data suggest an generally depicted in Figure 1 and illustrated specifically for effectively planar structure for the parent oxetane14 with a Figure 8. Change in the pK value of an amine upon replacement of a a cationsfunctional arising§ groups.Solubility from basic When! lipophilic situatedcyclic compounds close ketoamines to a forming basic or lactams amino in Figure 2, can provide an low pucker-inversion barrier. Substitution at the 3-position methylene group with an oxetane ring (difference in brackets). micellesgroup, at a the neutral observed pH value, gain intrinsic in solubilityinteresting solubilities may structural be at highmoderate alternative. leads to increased eclipsing interactions with the adjacent pK values are amine basicities measured spectrophotometrically at – Increases soluble up to 4000x than a gem-dimethyl derivative a pH valuesbecause and of lipophilicitiesthe reduction of (log basicityP) for and theAqueous concomitant neutral solubility amines of increase lipophilic scaffolds is often improved methylene groups, and therefore, more pronounced pucker- 248C in H2O. by the attachment of a morpholine unit, even though morpho- ing of the oxetane ring is found in many of these structures. in the intrinsic[2a,b]– lipophilicityChanges (Figure polarity 10; 60of andmolecules106). However, were determined. As demonstrated inline Figure is known 9,there to be a is likely a target of oxidative metabolism. Nevertheless, the puckering angle remains quite small, which significant decrease in lipophilicity onThere going are 17 from marketedgem- drug substances that contain the is important with regard to the structural analogy between an dimethyl compound 112 to its oxetanemorpholine derivative moiety. For62 13 of them metabolic data have been oxetane and a carbonyl unit. cations arising from basic lipophilic compounds forming published,10 and in 8 compounds oxidative degradation of An oxetane unit exerts a distinct influence on the con- (DlogP = 2.4), but the corresponding amino ketone 108 is 11 micelles at a neutral pH value, intrinsic solubilities at high À the morpholine ring (Figure 2) is the predominant meta- formational preferences of the backbone chain to which it is even less lipophilic by 0.4 logarithmicbolic units. pathway. Whereas The com- class of spirocyclic oxetanes shown in grafted (Figure 4). Whereas antiperiplanar conformations pH values and lipophilicities (logP) for the neutral aminespounds 64 and 109 differ by more thanFigure 1logP 2 mayunit, be there taken as is structural analogues of morpholine are favored for an unsubstituted aliphatic chain (τ = 0), all [2a,b] were determined. As demonstrated in Figure 9, there is aonly a small difference in the intrinsicin lipophilicity their ability to between increase solubility while maintaining stability three staggered conformations are approximately equally toward oxidative metabolism. However, these oxetane deri- populated at the site of a gem-dimethyl group (τ =0 , 120 , significant decrease in lipophilicity on going from gemoxetane- and b-lactam . Angew. Chem. Int. Ed. 2010, 49, 9052.! ° ° 66 110 vatives are intrinsically interesting building blocks through and -120°). For a 3,3-disubstituted oxetane, the synclinal dimethyl compound 112 to its oxetane derivative 62 All the available evidence taken togethertheir distinct shows spatial that distribution an of polarity. (gauche) conformations (τ = 120° and -120°) are clearly (DlogP = 2.4), but the corresponding amino ketone 108 isoxetane-containing molecule is typically much less lipophilic preferred over an antiperiplanar arrangement of the chain. À This is likely a result of the small oxetane C-C-C valence even less lipophilic by 0.4 logarithmic units. Whereas com-than the respective gem-dimethyl analogue,Results and and Discussion that the pounds 64 and 109 differ by more than 1logP unit, there is angle ( 84°) that leads to an increase of steric repulsion in Structure and Conformation of Oxetanes. The analogies the antiplanar∼ backbone arrangement compared to that of a only a small difference in the intrinsic lipophilicity between 15 regarding 3-substituted oxetanesFigure depicted 10. Change in Figure in 1 the rely aqueous on solubilitygem-dimethyl at pH group. 9.9 upon substitu- oxetane 66 and b-lactam 110. limited structural knowledge12 available from the Cambridge The analogy of an oxetane with a gem-dimethyl group 13tion with an oxetane unit. Sol.= intrinsic thermodynamic solubility All the available evidence taken together shows that an Structural Database (CSD). Additional1 X-ray structures draws on the experimental fact that the partial molar volume (mgmLÀ ) in 50 mm phosphate buffer measured at pH 9.9 and 3 17 were obtained in the course of this work highlighting oxetane of oxetane in water (61.4 cm /mol at 25 °C) is comparable oxetane-containing molecule is typically much less lipophilic (22.5 1)8C. than the respective gem-dimethyl analogue, and that the Æ

Figure 10. Change in the aqueous solubility at pH 9.9 upon substitu- if introduced more remotely, the oxetane may increase the tion with an oxetane unit. Sol.= intrinsic thermodynamic solubility (intrinsic) solubility by several orders of magnitude (see 116). 1 (mgmLÀ ) in 50 mm phosphate buffer measured at pH 9.9 and It is interesting to note that there is not a linear relationship (22.5 1)8C. Æ between the solubility and the distance of the basic amine group from the oxetane, as substitution in the g position (58) led to a higher solubility than oxetanes attached at the b or if introduced more remotely, the oxetane may increase thed positions. (intrinsic) solubility by several orders of magnitude (see 116). It is interesting to note that there is not a linear relationship between the solubility and the distance of the basic amine5.5. Metabolic and Chemical Stability group from the oxetane, as substitution in the g position (58) led to a higher solubility than oxetanes attached at the b or Many promising lead structures do not reach the clinical d positions. phase because of their metabolic instability. Several biological pathways are subsumed under the term metabolic degrada- tion.[75] The influence an oxetane has on oxidative phase I 5.5. Metabolic and Chemical Stability metabolism has been studied and compared with compounds having carbonyl or gem-dimethyl groups at the corresponding Many promising lead structures do not reach the clinicalposition.[2a,b] It was found that compounds bearing oxetanes Figurephase 9. Change because in lipophilicity of their metabolic upon exchange instability. of a carbonyl Several or a biological gem-dimethyl group with an oxetane ring. logP =intrinsic lipophilicity tend to have a low to modest proclivity towards metabolic pathways are subsumed under the term metabolic degrada-degradation, and that in many cases the corresponding gem- of the neutral[75] base. n.d. =not determined because of chemical instability.tion. The influence an oxetane has on oxidative phasedimethyl I or carbonyl compounds have higher intrinsic metabolism has been studied and compared with compounds 9062 www.angewandte.orghaving carbonyl or gem-dimethyl 2010 groups Wiley-VCH at the Verlag corresponding GmbH & Co. KGaA, Weinheim Angew. Chem. Int. Ed. 2010, 49, 9052 – 9067 position.[2a,b] It was found that compounds bearing oxetanes Figure 9. Change in lipophilicity upon exchange of a carbonyl or a gem-dimethyl group with an oxetane ring. logP =intrinsic lipophilicity tend to have a low to modest proclivity towards metabolic of the neutral base. n.d. =not determined because of chemical degradation, and that in many cases the corresponding gem- instability. dimethyl or carbonyl compounds have higher intrinsic

9062 www.angewandte.org  2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Angew. Chem. Int. Ed. 2010, 49, 9052 – 9067 Article Journal of Medicinal Chemistry, 2010, Vol. 53, No. 8 3235

Comparison of the hydrogen bonding avidity of oxetane contain an extra carbon atom compared to morpholine. with various carbonyl compounds demonstrates that oxe- The pyrrolidine analogues (18, 21)areintermediatewith tanes can compete favorably with aliphatic ketones, alde- intrinsic lipophilicities quite similar to that of the morpholine hydes, or esters27 (Figure 10), which is of interest in view of a derivative. For purposes of comparison we include here the similar spatial disposition of the oxygen lone pairs as in a (3,5)-methano-bridged morpholine analogue 32 which also carbonyl unit (Figure 1, middle). However, an oxetane is a exhibits very similar lipophilicity as its parent morpholine much weaker hydrogen bond acceptor than an amide carbo- derivative. nyl group (Figure 11). This is important to remember, apart For log D, the basicity-reducing factor of the oxetane from structural or basicity aspects (see above), when con- module adds to the pattern seen for log P, resulting in sidering the replacement of an amide or lactam by a 3-amino- dramatic polarity increases for the more basic spirocyclic oxetane unit.38 amines with the oxetane unit in β- or γ-position to the An oxetane analogue is typically more lipophilic than its nitrogen. Thus, the lipophilicity (log D)oftheβ-spiro- carbonyl counterpart because of the presence of two methyl- oxetane analogue 27 is comparable to that of the morpholine ene groups in the four-membered ring. The effect is most derivative 33 (Figure 13 right, blue bar in β-group), whereas pronounced for β-ketopiperidine and β-ketopyrrolidine, the γ-analogue 30 is considerably more polar. Moreover, the which are both unusually polar, possibly as a consequence β-spiro-oxetanes in the pyrrolidine and azetidine series, of the (1,3)-syn juxtaposition of the polar carbonyl and compounds 21 and 15 (Figure 13 right, red and green bars, amine or protonated amine functions with dipolar enhance- respectively, in the β-group), are significantly less lipophilic ment as well as possibilities for cooperative solvation. By than the morpholine counterpart. contrast, the amine and ether oxygen functions are insulated By virtue of its polarity, the morpholine unit often helps to from each other by an intervening methylene group of the improve the solubility of a compound into which it is oxetane ring in the β-spirocyclic analogues, as illustrated in integrated. Remarkably, the N-piperonylmorpholine deri- Figure 12. vative 33 has the highest solubility (36000 μmol/L) of all In view of the structural analogies between the spirocyclic compounds without an oxetane in this study. When morpho- Chemical Propertiesoxetane amines and morpholine, a direct comparison of their line derivative 33 is compared with its spirocyclic analogues lipophilicities and solubilities is instructive (Figures 13 and 14). (Figure 14, right), it is found to be more than twice as soluble. Although intrinsic lipophilicity is observed to increase in all This result is remarkable in view of the higher polarity of at oxetane derivatives of piperidine (24, 27, 30), the azetidine least some of the spirocyclic oxetane amine derivatives. The § Oxetanes have mostanalogues lewis (12 basic, 15)aremorepolarthanthecorresponding oxygen of cyclic ethers (pKa = 2-4)“homospiro-morpholine” 15 is one notable exception (green § H-bond acceptormorpholine derivative (33, green bars in Figure 13), despite bar in the β-group), however, as already noted above. This the fact that the spiro-oxetane derivatives of azetidine highlights the “homospiro-morpholine” as a very promising – Compete with aldehydes/ketones/esters solubilizing unit. In Figure 14 (left) we include a comparison of the solubi- lities of the spirocyclic oxetane amines with the correspond- ing cyclic ketoamines or lactams. It illustrates that all but one carbonyl derivative in this study are more soluble than the more lipophilic oxetane analogues. The exception is the case of the spirocyclic R-oxetanepyrrolidine derivative 18 (red bar in the R-group) that is slightly more soluble than its γ-butyrolactam counterpart 19.Unfortunately,theazetidine- 3-one 16 could not be included in this study because of its Minireviews inherent instability under various analyticalM. conditions, Rogers-Evans thus et al. precluding a direct comparison with the highly soluble β-oxetane amine analogue 15. § When substitutedFigure on a 10. molecule,Hydrogen bond consider acceptor aviditythem for as oxetane electron and dif- withdrawingChemical groups Stability. All 3,3-disubstituted oxetanes described ferent carbonyl compounds as measured by H-bonded complex correspondinghere were found carbonyl to be stable compounds when subjected are to even treatment more hydro- 37 formation in CCl4. philic.in aqueous Thus, solutions, the oxetane buffered positions at pH 1- itself10, for between 2 h at 37 ° aC carbonyl and a gem-dimethyl group, but is somewhat closer to the former. The observed trend in lipophilicity manifests itself in intrinsic aqueous solubilities.[2a,b] A carbonyl compound, under the same experimental conditions, is usually more soluble than its oxetane or gem-dimethyl analogues. The changes seen in the series 58, 60, 105,–107, and 116 emphasize the effects the introduction of the polar ethereal oxygen atom canAngew have,. Chem. especially Int. Ed. 2010 when, 49, 9052. employed! away from other polar

Figure 8. Change in the pKa value of an amine upon replacement of a functional groups. When situated close to a basic amino methylene group with an oxetane ring (difference in brackets). group, the observed gain in solubility may be moderate pKa values are amine basicities measured spectrophotometrically at because of the reduction of basicity and concomitant increase Figure 11. Changes in log P (A) and log D (B) upon replacement of a carbonyl group by an oxetane, as obtained from appropriate log P and 248C in H2O. log D value pairs in Table 1. Color codes of the bars are blue, red, andin green the for intrinsic the six-, five-, lipophilicity and four-membered (Figure cyclic 10; cases,60 and respectively.106). However,

cations arising from basic lipophilic compounds forming micelles at a neutral pH value, intrinsic solubilities at high pH values and lipophilicities (logP) for the neutral amines were determined.[2a,b] As demonstrated in Figure 9, there is a significant decrease in lipophilicity on going from gem- dimethyl compound 112 to its oxetane derivative 62 (DlogP = 2.4), but the corresponding amino ketone 108 is À even less lipophilic by 0.4 logarithmic units. Whereas com- pounds 64 and 109 differ by more than 1logP unit, there is only a small difference in the intrinsic lipophilicity between oxetane 66 and b-lactam 110. All the available evidence taken together shows that an oxetane-containing molecule is typically much less lipophilic than the respective gem-dimethyl analogue, and that the

Figure 10. Change in the aqueous solubility at pH 9.9 upon substitu- tion with an oxetane unit. Sol.= intrinsic thermodynamic solubility 1 (mgmLÀ ) in 50 mm phosphate buffer measured at pH 9.9 and (22.5 1)8C. Æ

if introduced more remotely, the oxetane may increase the (intrinsic) solubility by several orders of magnitude (see 116). It is interesting to note that there is not a linear relationship between the solubility and the distance of the basic amine group from the oxetane, as substitution in the g position (58) led to a higher solubility than oxetanes attached at the b or d positions.

5.5. Metabolic and Chemical Stability

Many promising lead structures do not reach the clinical phase because of their metabolic instability. Several biological pathways are subsumed under the term metabolic degrada- tion.[75] The influence an oxetane has on oxidative phase I metabolism has been studied and compared with compounds having carbonyl or gem-dimethyl groups at the corresponding position.[2a,b] It was found that compounds bearing oxetanes Figure 9. Change in lipophilicity upon exchange of a carbonyl or a gem-dimethyl group with an oxetane ring. logP =intrinsic lipophilicity tend to have a low to modest proclivity towards metabolic of the neutral base. n.d. =not determined because of chemical degradation, and that in many cases the corresponding gem- instability. dimethyl or carbonyl compounds have higher intrinsic

9062 www.angewandte.org  2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Angew. Chem. Int. Ed. 2010, 49, 9052 – 9067 Stability under Common Chemical Practices

§ Stable under basic conditions! – Ring opening very slow – LAH requires high temperatures and long reaction times to reduce ring – Organolithium/grignards require elevated temperatures and lewis acids to open

§ Acidic conditions! – Non-disubstituted oxetanes are stable above pH 1 – 3,3-disubstituted oxetanes stable even at pH 1 – Concentrated acid is problematic

• Acid-cat ring opening in dioxane with H2SO4 or HClO4 as fast as ethylene oxide – Strong Lewis acids coordinate well to promote transformations

§ Alkaline and weak acid stability allows oxetanes to be introduced early on in synthetic routes!

J. Med. Chem. 2010, 53, 3227.! Synthetic Routes to Oxetanes

§ First discovered in 1878 by Reboul

aq. base O HO Cl

Ann. Chim. 1878, 14, 496.!

§ Most Common Routes: 1. Intramolecular Williamson-Ether synthesis

R Base R X OH O 2. Sulfonium ylides to aldehydes O Me S CH Na O 2 NpTs O R R' R R' 3. Paternό-Bϋchi cycloaddition R O O R light R H + R R R

Angew. Chem. Int. Ed. 2010, 49, 9052.! 1. Intramolecular Williamson-Ether synthesis

R Base R X OH O

§ Most general approach – Most widely used and applicable

§ Difficult to predict substrate success – Chloro/bromo substrates can differ greatly

§ By-product formation is plentiful and hard to inhibit R R k1 Intra- O R R k R 2 + CO X OH Grob Fragmentation R

k3 HO O X Inter- R R R R J. Med. Chem. 2010, 53, 3227.! 546 F. W. Ng et al. / Tetrahedron Letters 43 (2002) 545–548

H. Lin et al. / Tetrahedron Letters 43 (2002) 549–551 551

Scheme 1. Synthetic plan.

(±)-gelsemine (1) is described. While the proposal pre- hyde 10.6 o-Nitrobenzylidenation7 of this aldehyde, sented above was realized in broad terms, reduction to using phosphonate 11, led to 13 presumably via divinyl- practice1. involvedIntramolecular exposure to many interesting issues Williamson-Ether in cyclopropane 12. It was envisioned synthesis that ketone 15 organic synthesis. would be an attractive type of intermediate to construct the critical oxetane moiety (cf. 5). The synthesis commenced with the epoxidation of 7-t- 5 butoxynorbornadiene§ (Scheme 2). Alumina promoted In principle, 15 could be readily obtained from alcohol rearrangement Danishefsky of epoxide 9 afforded Total Synthesis the known alde- of Gelsemine14, which might be reached by hydroboration–oxida-

Scheme 4. Reagents and conditions: (a) DIBAL, CH2Cl2, −78°C; (b) TsOH·H2O, CH2Cl2, reflux, 50% for two steps; (c) OsO4, THF, −25°C; Na2SO3 (aq.); (d) NaIO4,THF/H2O, 45%; (e) TESOTf, Et3N, CH2Cl2,0°C, 80%; (f) NaOMe, MeOH; (g) TPAP,

NMO, CH2Cl2,4A, MS, 50% for two steps; (h) TBAF/HOAc 1:1, THF, 80%; (i) Hg(OTf)2·C6H5NMe2,CH3NO2;NaBH4, 10%

NaOH, Et3BnNCl, CH2Cl2, 60%; (j) 10% NaOH, THF, 90%; (k) LiAlH4,THF,0–25°C, 81%. DIBAL=diisobutylaluminum hydride, TESOTf=triethylsilyltrifluoromethanesulfonate, TPAP=tetrapropylammonium perruthenate, NMO=N-methylmorpho- line N-oxide. Unfortunately, the focused gelsemine target goal 7. Cf: Winstein, S.; Buckles, R. E. J. Am. Chem. Soc. 1942, became quite complicated in that its solution required 64, 2787–2790. excision of a one carbon unit from a six-membered 8. Ng, F. W. Ph.D. Thesis, Columbia University, 1997, 54. lactam to a five-membered spiroanilide (see 13“19). In The stereochemistry at C14 of the rearrangement product

Scheme 2. Synthesisthe end, of the this oxetane ring ring. contractionReagents and was conditions accomplished.: (a) 11,NaOMe,DMF,0 A full °C, 74%;( (b)12) BH was2Cl· notDMS, determined. Et2O, 0°C; / 8 NaOH H2O2account, 77%, +7% regioisomer; of these(c) experiments (COCl)2,DMSO,Et and3N, other CH2Cl2 interesting, 98.7%; (d) LiHMDS,9. TESCl, Cf: EtStill,3N, THF, C. W.;−78 Mitra, to 0°C; A. J. Am. Chem. Soc. 1978, 100, Eschenmoser’s salt, CH Cl , 91%; (e) MeI, CH Cl /Et O; Al O ,CHCl , 95%; (f) NaBH ,CeCl·7H O, MeOH, 99%; (g) 9-BBN excursions2 2 directed to gelsemine2 2 2 (1)2 is3 planned.2 2 4 3 1927.2 dimer, THF; NaOH/H2O2, 88%; (h) MsCl, Et3N, CH2Cl2, −78°C; NaHMDS, THF, −78°C, 91%. DMS=dimethyl sulfide;

HMDS=hexamethyldisilazane; TESCl=chlorotriethylsilane; Eschenmoser’s salt=(CH3)2N!CH10.2I; 9-BBN For an=9-borabicyclo[3.3.1]- overview of [2,3]-Wittig rearrangements see: (a) nonane. Mikami, K.; Nakai, T. Org. React. 1994, 46, 105; (b) Nakai, Acknowledgements T.; Mikami, K. Chem. Rev. 1986, 86, 885–902. 11. Felix, D.; Gschwend-Steen, K.; Wick, A. E.; Eschenmoser, This work was supported by grants from the National A. Helv.DanishefskyChim. Acta. Tett1969. Lett, .52 2002, 1030, 43,– 545.1042.! Institute of Health (grant HL25848). H.L. would like to 12. Compound 14 was prepared to clarify the stereo outcome thank the Texaco Foundation for a postdoctoral fel- of the Claisen rearrangement as shown in Scheme 3. The lowship. We thank Dr. George Sukenick and Ms. Sylvi most conclusive signal was the NOE between one of the Rusli of the MSKCC NMR Core Facility for NMR a-protons in the lactam and H19 (the methine proton of and MS spectral analyses (NIH Grant CA08748). the terminal alkene). 13. Nishizawa, M.; Takenaka, H.; Hayashi, Y. J. Org. Chem. 1986, 51, 806–813. References 14. Newcombe, N. J.; Ya, F.; Vijn, R. J.; Hiemstra, H.; Speckamp, W. N. J. Chem. Soc., Chem. Commun. 1994, 1. Ng, F. W.; Lin, H.; Tan, Q.; Danishefsky, S. J. Tetrahedron 767–768. Lett. 2002, 43, 545–548. 15. Fukuyama, T.; Liu, G. J. Am. Chem. Soc. 1996, 118, 2. Saxton, J. E. In The Alkaloids; Manske, R. H. F., Ed.; 7426–7427. Academic Press: New York, 1965; Vol. 8, pp. 93–117. 16. Reductive demercuration was capricious when conducted 3. For specific examples of this effect see: Ng, F.W. Ph.D. Thesis, Columbia Unversity, 1997, 70, and 74. on small scale. It is currently under optimization. Hydroly- 4. For free radical benzylic bromination of substituted nitro- sis of 21 was performed on the material that was degraded toluene derivatives with NBS, see: Mataka, S.; Kurisu, M.; from commercially available gelsemine (1) in three steps: Takahashi, K.; Tashiro, M. Chem. Lett. 1984, 1969–1972. demethylation of 1 with PhOCOCl and Hu¨nig’s base 5. Following precedents, methylene chloride was used to provided the phenyl carbamate, which was converted to minimize double bromination, see: Offermann, W.; Vogtle, methyl carbamate (which is identical to 22)withNaOMe F. Angew. Chem., Int. Ed. Engl. 1980, 19, 464–465. in MeOH at reflux. Protection of the free oxindole with 6. Nakamura, E.; Inubushi, T.; Aoki, S.; Machii, D. J. Am. CbzCl, Et3NandDMAPinCH2Cl2 provided 21 whose Chem. Soc. 1991, 113, 8980–8982. For additional examples spectroscopic and chromatographic properties were identi- of free radical initiated oxygenation, see: Mayer, S.; Prandi, cal to those of its synthetic counterpart. J. Tetrahedron Lett. 1996, 37, 3117–3120; Moutel, S.; 17. Confalone, P. N.; Huie, E. M. J. Org. Chem. 1985, 52, 79–83. Prandi, J. Tetrahedron Lett. 1994, 35, 8163–8166. The yield 18. Ng, F. W.; Chiu, P.; Danishefsky, S. J. Tetrahedron Lett. was based on the recovered starting material. 1998, 39, 767–770. F. W. Ng et al. / Tetrahedron Letters 43 (2002) 545–548 547 tion of 13. Such a hydroboration reaction raised signifi- was converted by reduction to its allylic alcohol coun- cant questions of chemoselectivity among the two dou- terpart (23 and 24, respectively).15 These isomers were ble bonds, regioselectivity at the C5!C16 olefin and face individually treated with triethylorthoacetate as selectivity. The key issue was that of regiopreference, shown.16 Remarkably, each allylic alcohol gave rise to a even assuming, as we did, that reaction would be single and identical g,d-unsaturated ester 26 (pre- directed to the more strained and more exposed sumably via 25) with the b-vinyl and a-carboxymethyl cyclopentene (C5!C16) linkage. Our findings and con- functions at C20 in the required sense.17 This stereo- jectures on this kind of hydroboration as well as related chemical convergence might arise from the tendency of reactions, in a model closely related to 13, have been the enolate like component of the Claisen rearrange- discussed elsewhere.8 In the event, treatment of 13 with ment step to glide over the five-membered ring fused to

BH2Cl·DMS followed by oxidative workup, as shown, oxetane (see 25“26). Additional cases must be evalu- afforded a 11:1 ratio of alcohol 14 with the newly ated to distinguish between possible steric or electronic introduced alcohol at C5, relative to its isomer where factors in directing the face of the migration step. the alcohol is at C16. Oxidation9 of 14 afforded ketone 15. Regardless of the reasons for this convergence, it pro- vided smooth access to a key intermediate, 26. Alkaline The campaign to install the oxetane commenced with hydrolysis of the ethyl ester function served to release dimethylaminomethylation of the silyl enol ether the free acid 27.18 Subjection of the latter to Curtius derived from 15.10 Following quaternization of the degradation, as practiced by Shiori, afforded urethane nitrogen, and base induced elimination, the a- 28.19 As anticipated, the hitherto robust oxetane link- methyleneketone 16 was in hand. At this stage we could age, which had survived in the sequence that started take advantage of i-face addition to both sp 2 centers (C5 with 19, was opened by the urethane nitrogen under H. Lin et al. / Tetrahedron Letters 43 (2002) 54920 –551 551 and C16). Hydride delivery at C5, in the context of a Lewis acid activation (BF3 etherate) and compound Luche reaction,11 afforded 17. Hydroboration of 17 29 was in hand. also occurred from the b-face generating diol 18.12 From this diol intermediate, the a-face oxetane (19) was In summary, we have shown the viability of a synthetic fashioned in a straightforward way as shown. strategy organized around the central idea of using an oxetane linkage to store molecular functionality in a With the critical oxetane in hand, we entered the next compact setting. In this case, the logic was used to phase of the projected plan hoping to reach a func- deliver a highly hindered hydroxymethyl function. Key tional1. versionIntramolecular of allylic alcohol 5. We were anticipating Williamson-Etherselectivity issues with potentially broadersynthesis ramifications a [3,3]-type rearrangement en route to structure type 6. in synthesis were resolved favorably in the regioselective Following the cleavage of t-butyl ether 19,13 the result- hydroboration of 13 and in the stereoconvergent rear- ing alcohol function in 20 was oxidized to afford ketone rangements of 23 and 24“26. The progression of 29 to 21 (Scheme 3). Emmons-type condensation14 was suc- gelsemine, requiring responses to some difficult and cessful§ in termsOxetane of overall moiety yield, butused led toto a store 3:2 mixture molecularunanticipated functionality challenges, is described in an accompany- of b,b-disubstituted steroisomers 22. Each compound ing paper.

Scheme 4. Reagents and conditions: (a) DIBAL, CH2Cl2, −78°C; (b) TsOH·H2O, CH2Cl2, reflux, 50% for two steps; (c) OsO4, THF, −25°C; Na2SO3 (aq.); (d) NaIO4,THF/H2O, 45%; (e) TESOTf, Et3N, CH2Cl2,0°C, 80%; (f) NaOMe, MeOH; (g) TPAP,

NMO, CH2Cl2,4A, MS, 50% for two steps; (h) TBAF/HOAc 1:1, THF, 80%; (i) Hg(OTf)2·C6H5NMe2,CH3NO2;NaBH4, 10%

NaOH, Et3BnNCl, CH2Cl2, 60%; (j) 10% NaOH, THF, 90%; (k) LiAlH4,THF,0–25°C, 81%. DIBAL=diisobutylaluminum hydride, TESOTf=triethylsilyltrifluoromethanesulfonate,30 = TPAPkey intermediate=tetrapropylammonium perruthenate, NMO=N-methylmorpho- line N-oxide. Unfortunately, the focused gelsemine target-19 more goal steps 7.to Cf:final Winstein, prdocut S.; Buckles, R. E. J. Am. Chem. Soc. 1942, became quite complicated in that its solution required 64, 2787–2790. excision of a one carbon unit from a six-membered 8. Ng, F. W. Ph.D. Thesis, Columbia University, 1997, 54. lactam to a five-membered spiroanilide (see 13 19). In Scheme 3. Construction of quaternary C7 and the pyrollidine ring. Reagents“ and conditions:(a)TFAThe stereochemistry/CH Cl ,0°C, 81%; at C14 (b) of the rearrangement product the end, this ring contraction was accomplished. A full 2 2 (COCl)2,DMSO,Et3N, CH2Cl2, −78°C, 81%; (c) triethylphosphonoacetate, NaH, THF, 0°C,( 3:2,12) 92%; was (d) not DIBAL, determined. CH2Cl2, −78°C, 88%;account (e) cat. propionic of these acid, H experiments3CC(OEt)3, toluene, and reflux, other 64%; interesting(f) NaOH/THF/EtOH,9. 86%; Cf: (g) Still, diphenylphosphoryl C. W.; Mitra, azide, A. J. Am. Chem. Soc. 1978, 100, Et3N, benzene,excursions 25°C, refl directedux; MeOH, to refl gelsemineux; 89%; (h) (BF1)3·Et is2O, planned. CH2Cl2, −78 to 12°C, 64%; (i)1927. PivCl, Et3N, DMAP, CH2Cl2, = = = 0–25°C, 92%. DIBAL diisobutylaluminum hydride; PivCl 2,2,2-trimethylacetyl chloride;10. DMAP For anN overview,N-dimethylaminopy- of [2,3]-Wittig rearrangements see: (a) ridine. Mikami, K.; Nakai, T. Org. React. 1994, 46, 105; (b) Nakai, Acknowledgements T.; Mikami, K. Chem. Rev. 1986, 86, 885–902. 11. Felix, D.; Gschwend-Steen, K.; Wick, A. E.; Eschenmoser, This work was supported by grants from the National A. Helv.DanishefskyChim. Acta. Tett1969. Lett, .52 2002, 1030, 43,– 545.1042.! Institute of Health (grant HL25848). H.L. would like to 12. Compound 14 was prepared to clarify the stereo outcome thank the Texaco Foundation for a postdoctoral fel- of the Claisen rearrangement as shown in Scheme 3. The lowship. We thank Dr. George Sukenick and Ms. Sylvi most conclusive signal was the NOE between one of the Rusli of the MSKCC NMR Core Facility for NMR a-protons in the lactam and H19 (the methine proton of and MS spectral analyses (NIH Grant CA08748). the terminal alkene). 13. Nishizawa, M.; Takenaka, H.; Hayashi, Y. J. Org. Chem. 1986, 51, 806–813. References 14. Newcombe, N. J.; Ya, F.; Vijn, R. J.; Hiemstra, H.; Speckamp, W. N. J. Chem. Soc., Chem. Commun. 1994, 1. Ng, F. W.; Lin, H.; Tan, Q.; Danishefsky, S. J. Tetrahedron 767–768. Lett. 2002, 43, 545–548. 15. Fukuyama, T.; Liu, G. J. Am. Chem. Soc. 1996, 118, 2. Saxton, J. E. In The Alkaloids; Manske, R. H. F., Ed.; 7426–7427. Academic Press: New York, 1965; Vol. 8, pp. 93–117. 16. Reductive demercuration was capricious when conducted 3. For specific examples of this effect see: Ng, F.W. Ph.D. Thesis, Columbia Unversity, 1997, 70, and 74. on small scale. It is currently under optimization. Hydroly- 4. For free radical benzylic bromination of substituted nitro- sis of 21 was performed on the material that was degraded toluene derivatives with NBS, see: Mataka, S.; Kurisu, M.; from commercially available gelsemine (1) in three steps: Takahashi, K.; Tashiro, M. Chem. Lett. 1984, 1969–1972. demethylation of 1 with PhOCOCl and Hu¨nig’s base 5. Following precedents, methylene chloride was used to provided the phenyl carbamate, which was converted to minimize double bromination, see: Offermann, W.; Vogtle, methyl carbamate (which is identical to 22)withNaOMe F. Angew. Chem., Int. Ed. Engl. 1980, 19, 464–465. in MeOH at reflux. Protection of the free oxindole with 6. Nakamura, E.; Inubushi, T.; Aoki, S.; Machii, D. J. Am. CbzCl, Et3NandDMAPinCH2Cl2 provided 21 whose Chem. Soc. 1991, 113, 8980–8982. For additional examples spectroscopic and chromatographic properties were identi- of free radical initiated oxygenation, see: Mayer, S.; Prandi, cal to those of its synthetic counterpart. J. Tetrahedron Lett. 1996, 37, 3117–3120; Moutel, S.; 17. Confalone, P. N.; Huie, E. M. J. Org. Chem. 1985, 52, 79–83. Prandi, J. Tetrahedron Lett. 1994, 35, 8163–8166. The yield 18. Ng, F. W.; Chiu, P.; Danishefsky, S. J. Tetrahedron Lett. was based on the recovered starting material. 1998, 39, 767–770. 2. Sulfonium Ylides

§ One-pot conversion of aldehydes/ketones to 2-substited oxetanes

§ Variant of Corey-Chaykovsky Reaction

O

Me S CH2Na NpTs Proposed Mechanism O (2.5 equiv.) O R R' DMSO, r.t. R R' O NaH2C S Me O O NpTs (1st equiv.) -O O O S NpTs O 47% R R' R R' O

NaH2C S Me SO2NpTs O - O O NpTs O 59% R R' (2nd equiv.) R R'

O O O 79% R R' H H

Welch. Rao. JACS 1979, 101, 6135.! Minireviews M. Rogers-Evans et al. 2. Sulfonium Ylides 2,2-disubstituted oxetane products are formed in good yield and selectivity.[38] For example, the treatment of camphor (20) with three equivalents of 21 in DMSO afforded oxetane 22 in 59%§ yield.Start The with reaction epoxides of 21 with to aform collection more of complicated other oxetanes aliphatic ketones afforded the corresponding oxetanes in Me similar yields. OEt H O, HOAc TsCl, pyr R O 2 OH OkumaO et al. showedKOtBu, that Me3 dimethyloxosulfoniumSOI O methyl-O r.t. ON O 0 C, ON R ° ide (27) can be usedtBuOH, in the50 °C conversion of ketones into 74% 73% oxetanes.[39] Following56-83 these% findings and their previous studies on theR = asymmetricCH2OH, C2H4CH=CH Corey–Chaykovsky2 reaction, Cl SNa Cl Scheme 1. Fundamental synthetic pathways towards oxetanes. Shibasaki and co-workers developed an asymmetric synthesis OTs O very tolerant O of 2,2-disubstituted oxetanes (Scheme 4).[40] The starting MeOH, r.t. ON S Fitton. Synthesis 1987, 12, 1140.! as the Paternò–Büchi reaction.[34] Both transformations have 74% been extensively discussed;[35] consequently this Minireview will only highlight variants that are general and versatile. Chiral Version! ! 3.1. Stereoselective Syntheses § Can work by direct conversion but lower

The enantioselective synthesis of substituted oxetanes is a yields particular challenge. Soai et al. documented an efficient approach for the preparation of optically active 2-substituted § 2 steps: increases rate of epoxide oxetanes (Scheme 2). The enantioselective reduction of a b- opening reaction

§ 99.5% ee (2nd step works as resolution process)

Shibasaki. Angew. Chem. Int. Ed. 2009, 48, 1677.!

Scheme 4. One-pot sequential addition of sulfur ylides to ketones to form 2,2-disubstitued oxetanes according to Shibasaki and co-work- ers.[40] M.S.= molecular sieves. Scheme 2. Enantioselective synthesis of 2-aryloxetanes by Soai et al.[36] methyl ketones 23 were treated with dimethyloxosulfonium haloketone followed by ring closure furnishes the targeted methylide (27, 1.2 equiv) and catalytic amounts of 26 [36] heterocycle. For example, treatment of 3-chloropropiophe- (5 mol%) in the presence of an equal amount of a Ar3P=O none (17) with a catalyst prepared in situ from LiBH4 and additive (28) and molecular sieves to give the corresponding cystine-derived chiral ligand (R,R)-19 afforded the corre- epoxides 24. The low reaction rate in the subsequent epoxide sponding optically active secondary alcohol, which is sub- opening reaction was compensated by the addition of another sequently subjected to cyclization to give 2-phenyloxetane equivalent of ylide and 15 mol% of catalyst in addition to (18) with 84% ee. heating. The one-pot sequential addition of a sulfur ylide to Oxetanes have been prepared directly from ketones via methyl ketones furnished 2,2-disubstituted oxetanes 25 in intermediate epoxides through the implementation of an good yield and up to > 99.5% ee. A resolution process was intriguing variant of the Corey–Chaykovsky reaction observed in the second step of the sequence, as the (Scheme 3). Following oxirane formation, excess ylide attacks intermediate epoxides were observed to be formed with the epoxide to give an oxydialkylsulfoximine intermediate,[37] lower enantioselectivity than the isolated oxetane products. which in turn undergoes displacement with ring closure. The The direct conversion of ketones into oxetanes was also possible at 458C by using 2.2 equivalents of ylide and 20 mol% each of the catalyst and additive. The Paternò–Büchi reaction offers rapid access to sub- stituted oxetanes. However, the control of facial selectivity in the [2+2] cycloaddition reaction is not easily realized. Recent studies by Bach et al. revealed that the photochemical reaction of benzaldehyde and silyl enol ethers can lead to the diastereoselective ring formation.[41] Thus, the cycloaddi- Scheme 3. Construction of 2,2-disubstituted oxetanes through sequen- tion reaction of alkoxy silyl enol ether 29 affords adducts as an tial additions of sulfur ylides. Ts = toluene-4-sulfonyl. 85:15 mixture of stereoisomers (Scheme 5). The diastereose-

9056 www.angewandte.org  2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Angew. Chem. Int. Ed. 2010, 49, 9052 – 9067 3. Paternό-Bϋchi Cycloaddition

§ Reaction first discovered by Emanuele Paternò in 1909 – Structure unconfirmed

O O O sunlight H + +

§ Re-examined in 1954 by George Büchi – extended reaction

O UV lamp O O + + Alk H Alk Alk

§ Reaction promoted by UV light – Carbonyl species is usually the light absorbing species

§ Access to substituted oxetanes

Paterno Gazz. Chim. Ital. 1909, 39, 341 Büchi. JACS 1954, 76, 4327. ! 3. Paternό-Bϋchi Cycloaddition 1. Introduction § Mechanism ISC

* * * O O O + + S 1 Exciplex T1

h!

O O O 1. IntroductionIS C +

12,23 24 S0 biradicals were trapped directly with oxygen and sulfur dioxide. Freilich and Peters have 1,4-S-BR 1,4-T-BR detected the 1,4-biradical from benzophenone and 1,4-dioxene which was also trapped by oxygen to give a 1,2,4-trioxane. Ph O O – O O Triplet 1,4-biradical O O2 O h! Ph + Ph O Ph Ph O O O O Ph " = 525 nm Oxetane # = 1.6 ns Scheme 1.2: Mechanism of Paternò-Büchi reactions.Scheme 1.3: Trapping of 1,4-triplet biradical with oxygen.

1.5 Lifetime of triplet 1,n-biradicals It is widely accepted, that the immediate precursors of the oxetanes are biradicals10 whose The lifetime of triplet 1,n-biradicals is usually determined by the intersystem crossing rate Koecky. Organic11 Photochemistry, A Visual Approach. VCH Publishers, New York 1992, 126.! existance has been evidenced by chemical means (# BRas = well 1/kISC as). Thisby the means, application that the biradical of picosecond has to stay in the triplet manifold for a defined spectroscopy.12 Two mechanisms have been proptime,osed because to wit explainhout spin the flip low it - canenergy not undergo pathways bond formation or bond cleavage. These 25 leading to the biradical intermediate.10,13 steps would lead to triplet excited open-shell species and thus would be highly endergonic. Three mechanisms operate for the interaction between singlet and triplet states of 1,n- * (1) Nucleophilic attack initiated by the half-occupiedbiradicals: " - electronorbital- nuclear of the hyperfine carbonyl coupling oxygen (HFC), atom spin -lattice relaxation (SLR), and spin- to the unoccupied "*-orbital of an electron-deficientorbit couplingolefin in (SOC). the 26 plane HFC ofis anthe important molecule. control This factor for biradicals with long carbon 27 LUMO-LUMO interaction is called “parallel approach”.chains between the radical centers. SOC plays the dominant role in biradicals with shorter distances between the radical centers, whereas SLR seems to contribute only marginally in (2) An electrophilic attack initiated by the halfgeneral. occupied In contrast n-orbital to other of mechanisms, the carbonyl SOC is oxygen strongly dependent on the geometry of the atom to the unoccupied "*-orbital of an electron triplet rich olefin biradical. in This a perpendicular was first summarized direction in three to rulesthe stated by Salem and Rowland28 in plane of the molecule. This HOMO-HOMO interaction1972: is called a “perpendicular approach”. (a) SOC decreases with increasing distance between the two spin-bearing atoms. Because Thus, the immediate precursors of these intermediates are 3 * or 3n * states. From their there is also the" " possibility " of through-bond interaction, not only is the actual distance interaction with a suitable ground state substrate, between three types the two of radical interme centersdiates important can possibly (corresponding be to a conformational dependence) formed:13 but also the number of bonds (n-1). (b) Conservation of total angular momentum could be achieved when the axes of the p (a) an exciplex with the excitation localized on one of the two partners. orbitals at the radical centers are oriented perpendicular to each other (see Figure1.4). (b) a neutral “conventional” biradical. (c) an ionic biradical or radical ion pair.

4 6 3. Paternό-Bϋchi Cycloaddition

§ Most stable diradicals favored, as well as stereochemical approach of radicals

O h! O + Ph Ph Me Me 93% Ph Ph

iPr iPr O h! O + Ph Ph SMe 73% Ph Ph SMe H O h! O + Ph H O 51% O Ph H § Reaction commonly employed for chiral product synthesis

MeO Me MeO Me Me Me H Me O O H h! O O OMe PhCHO H H OMe + OMe vs. tBu OTMS 30 °C Ph t H t tBu OTMS Bu OTMS Bu OTMS Ph OTMS Ph tBu H Ph 85 : 15 64% yield

Bach. Angew. Chem. Int. Ed. 1995, 34, 2271. Bach. JACS 1997, 119, 2437. ! Most Targeted Oxetane

§ Functionalization of oxetanes is difficult, therefore oxetan-3-one is common starting point

§ First synthesis by Marshall in 1952

O O O 1) CH2N2, Et2O - O Cl OH Cl Cl 2) K2CO3, MeOH/H2O O N2 OH N2

Marshall. J. Chem. Soc. 1952, 467. – Converted to hydrazone for isolation

§ Many procedures require prep GC to isolate pure § General route established by Carreira

MeO OMe O O HC(OMe)3 MeO OMe 1) BuLi, TsCl Montmorillonite K10 HO OH TsOH HO OH 2) NaH 62% yield 2 O O 3 steps: 33% yield

Muller, K.; Carreira, E.. Angew. Chem. Int. Ed. 2006, 118, 7900. ! Angewandte Oxetanes Chemie

generating a stereogenic center upon grafting an oxetane unit onto a given scaffold. In developing a general approach for the facile, versatile incorporation of 3-substituted oxetanes, we have concentrated on oxetane-3-one as a starting point. Several syntheses of oxetan-3-one (35) had been reported; however, in our assessment each has practical shortcomings, such as the use of preparative gas-chromatographic tech- niques.[45–47] Consequently, a straightforward route was devel- oped starting from dihydroxyacetone dimer (36; Scheme 7).

Scheme 5. Paternò–Büchi reaction between benzaldehyde and a silyl enol ether. TMS=trimethylsilyl.

lectivity arises from the inherent conformational preferences that arise from 1,3-allylic interactions with the trisubstituted olefin partner. Scheme 7. Synthesis of oxetan-3-one.

In this procedure, dihydroxyactone is converted into the 3.2. Elaboration of Oxetanes corresponding dimethyl ketal 37, which is subsequently treated with TsCl and subjected to ring closure.[2a] Acidic The preparation of 3-aryl-substituted oxetanes by means cleavage of the ketal 38 and distillation furnishes oxetan-3- of nickel-mediated Suzuki coupling reactions was recently one (35). reported (Scheme 6).[42] Treatment of 3-iodooxetane (33) with Oxetan-3-one is amenable to extensive functionalization and elaboration (Scheme 8).[47b,48] It undergoes addition Oxetan-3-onereactions with Applications organometallic reagents to yield oxetan-3-ols

Scheme 6. Nickel-mediated Suzuki coupling of 3-iodooxetane with arylboronic acids. NaHMDS= sodium hexamethyldisilazide, MW = microwaves.

arylboronic acids in the presence of 6 mol% of NiI2/trans-2- aminocyclohexanol hydrochloride[43] and NaHMDS (2 equiv) at 808C under microwave irradiation provides access to a wide range of substituted oxetanes 34. 3-Arylazetidines were shown to be accessible in the same fashion from N-Boc- protected (Boc = tert-butoxycarbonyl) 3-iodoazetidine. Alter- natively, similar heteroaryl oxetanes and azetidines can be readily accessed by a Minisci reaction.[44] The methods described above provide access to a range of substituted oxetanes. These will certainly be beneficial for a variety of applications.

3.3. Chemistry Based on Oxetan-3-one

The initial investigations on oxetanes as useful modules in Angew. Chem. Int. Ed. 2010, 49, 9052.! drug discovery largely focused on 3,3-disubstituted oxetanes. Scheme 8. Reactions of oxetan-3-one in the literature.[47b, 48] Z =benzyl- This stems from the aim of not augmenting the complexity by oxycarbonyl.

Angew. Chem. Int. Ed. 2010, 49, 9052 – 9067  2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.angewandte.org 9057 Oxetan-3-one226 ApplicationsProcess Chemistry in the Pharmaceutical Industry, Volume 2

TMSCl, CuI, THF, - CO2Et CO2Et BrMg 18 °C § Methylene compounds O (1) + 70% O tert-Bu seem very stable 8 § Applicable to add in tert-Bu late-stage

cat. [Rh(cod)Cl]2, CO2Et functionalization of (HO)2B CO2Et KOH, aq dioxane, RT + O (2)molecules 83% O NMe2 3 NMe2 8 3 O cat. [Rh(cod)Cl] , B(HO)2 2 CHO CHO KOH, aq dioxane, RT (3) O + tert-Bu 83% tert-Bu 9

NO cat. [Rh(cod)Cl] , NO2 2 (HO) B 2 2 KOH, aq dioxane, RT + tert-Bu O (4) 10 tert-Bu 51% O

Me2N CHO CHO (5) O HNMe2, DBU, O 9 THF, –18 °C Gadamasetti and Braish. Process Chemistry in the Pharmaceutical Industry 2008, Vol 2, CRC Press. SCHEME 13.7

oxetan-3-one has been suggested to undergo self-condensation or polymerization under basic conditions. Additionally, condensation of oxetan-3-one with nitromethane furnishes nitromethylene oxetane 10 in 81% yield. Surprisingly, none of these compounds has previously been reported, yet they constitute tractable, stable entities amenable to storage. The α,β-unsaturated ester (8), aldehyde (9), and nitro (10) compounds participate as electro- philes in a number of useful conjugate addition reactions. Copper-catalyzed addition of Grignard reagents provides access to aryl butanoic acid derivatives substituted with an oxetane (equation 1 in scheme 13.7). The ester, aldehyde, and nitro electrophiles also undergo mild Rh-catalyzed additions of aryl and vinyl boronic acids (equation 2 through equation 4 in scheme 13.7). Interest- ingly, the unsaturated aldehyde participates readily in an amine conjugate addition to afford oxetane substituted 3-amino-acetaldehyde derivatives.

13.7 CONCLUSION Improved access routes to relatively small building blocks can open up new opportunities in drug discovery. We detailed the case of oxetan-3-one; this simple ketone is rich in the chemical reactions it and its derivatives participate in. We anticipate additional studies focused on this building block, Methodology of Oxetanes

Ring Opening Organometallic Reacons Reacons Simple Acid and Base Ring Opening

H HO O MeOH, CSA Me Me H O Me

OH O LAH, THF R Me 1,2 syn diol Ph Ph OH R OH

Me H SiMe Ph OH 2 TBAF, THF O Me R Ph Ph OH R OH

Bach. Liebigs Ann. 1997, 1627. Bach. Synthesis 1998, 683. ! !"# !"#$%&'(#)"#*+, $%&'()*$

8':3A+:3B#;'#43&A;#3CC7A73,;:D#>7;%#7B'AD&,&;3B#-#7,#&#E-FGH OTIPS O C&B%7',#;'#C'48#43I+A3IJ%DI&,;'7,#BA&CC':I#.#E*A%383#G OTIPS HNTf2 N (10 mol%) KLMH" #O,BP743I#QD#?+I7,RB#B+AA3BB(#>3#%DP';%3B7S3I#;%&; + Ar CH Cl , r.t. O KT(,MJ&8P%';347A#8':3A+:3B#K,#UVM#8&D#+,I345'#ADA:7S&J O 2 2 Ar ;7',#>7;%#I7P':&4'P%7:3B#;'#C'48#83I7+8#&,I#:&453#47,5B" R O 12 13 Expansion!3#3,W7B7',3I#B;4+A;+43# to Medium,"(#>%7A%#7,A'4P'4&;3B#&,#&:I3J Sized Lactone Rings R %DI3#&,I#&,#'X3;&,3#+,7;#;'53;%34#>7;%#&#;>'J&;'8#:7,@J 14 34"0(Y#Z,A3#;%3#&:I3%DI3#54'+P#7B#&;;&A@3I#QD#& § Most approaches start from linear substrate and must overcome unfavorable OTIPSring closing OTIPS ,+A:3'P%7:3(#;%3#43B+:;7,5#,+A:– Requiring high dilution and 3'P%7:7A#'XD53,#&;'8#7B#3XJslow addition P3A;3I#;'#7,7;7&;3#'X3;&,3#'P3,7,5#&,I#53,34&;3#&,';%34O F ,+A:3'P%7:7A#'XD53,#&;'8#&;#;%3#NJP'B7;7'," [act] O O

OH O O O (A) Yudin's design:(1) (1,3)-amphotericCOOH molecules R (14e), 62% R OH O O R1 R: R: 1 R O O- n-Bu (14a), 71% n-Bu (14f), 70% 3 [3+2] t-Bu (14b), 60% t-Bu (14g), 87% NH N + C [act] n-Oct (14c), 64% n-Oct (14h), 74% 2 1 O Ph (14d), 47% Ph(CH ) (14i), 51% N X X 2 3 R2 N TIPSO(CH ) (14j), 71% O O RCM. olefinationHO O 2 3 (2) Y Y O radical cyclization, etcR2 O 7 89 OTIPS OTIPS § Amphoteric molecule could cyclize with dipolarophiles (B) Our desgin: (1,6)-amphoteric molecules MeO – (electrophilic and nucleophilic moieties) O O 5 6 S O MeO O O 4 [6+2] R 3 + R 2 1 O R: R: n-Bu (14k), 47% n-Bu (14m), 49% Sun, J. Angew. Chem. Int.t-Bu Ed. (201214l),, 5152%, 6209.! t-Bu (14n), 53% 10 11 Ph (14o), 48%

$/01213+ #L8P%';347A#8':3A+:3B#7,#47,5#C'48&;7',#43&A;7',B $/01213! #b75%;J838Q343I#:&A;',3#BD,;%3B7B#W7&#ENFGH#ADA:7S&;7',

6%343C'43(#;%3#B;4+A;+43#A'+:I#Q3#A',B7I343I#KT(NMJ&8P%'J ,-#W7&#+,"#!3#%&W3#I',3#B'83#83A%&,7B;7A#B;+I73B"#c'4 ;347A#&,I#7;#7B#P'BB7Q:3#C'4#7;#;'#+,I345'#ENFGH#ADA:7S&;7', 3X&8P:3(#>3#%&W3#BD,;%3B7S3I#&:@3,3#+"#Kd#e#]M#&,I#B+QJ >7;%# I7P':&4'P%7:3B# ;'# C'48# 83I7+8J47,5# 8':3A+:3B# ,, f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f'4#:787;&;7',(#7"3"(#&,#&4D: A',I7;7',B(# ;%3# 43&A;7',# 3X%7Q7;B# 5''I# C+,A;7',&:# 54'+P :7,@34#Q3;>33,#;%3#'X3;&,3#&,I#&:I3%DI3#8'73;73B#7,#;%3 A'8P&;7Q7:7;D"# *;347A&::D# %7,I343I# &:@D,3B# &43# &:B'# B+7;J &8P%';347A# 8':3A+:3B# 7B# 43`+743I# ;'# &A%73W3# 43&B',&Q:3 &Q:3#B+QB;4&;3B"#]3;34'ADA:3JC+B3I#:&A;',3B#K,#2/4M#A&, 3CC7A73,AD"#c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`+3,;#ADA:7S&;7',#Q3;>33,#;%3#438';3#%DI4'XD:#54'+P#&,I 83I7+8J47,5#:&A;',3#BD,;%3B7B(#P&4;7A+:&4:D#;'#&II43BB#;%3

;%3#@3;3,7+8#8'73;D#C'48B#ADA:7A#B7:D:#@3;3,3#&A3;&:#,7" &Q'W3# B+QB;4&;3# :787;&;7',(# >3#%&W3#43A3,;:D#43P'4;3I#& Downloaded by: IP-Proxy University of Texas at Austin, Austin Libraries. Copyrighted material. =3X;(#&A;7W&;7',#'C#,7#QD#P4';',&;7',#'C#;%3#3;%34#8'73;D A',A3P;+&::D# I7CC343,;# B;4&;35D# A%&4&A;347S3I# QD# &,# +,J ;475534B#;%3#a௅Z#Q',I#A:3&W&53#;'#C'48#:&A;',3#,-(#>%7A% P43A3I3,;3I# 47,5# 3XP&,B7',# 'C# Q7ADA:7A# 'X3;3,7+8# +# 7B#B7:D:&;3I#;'#I3:7W34#;%3#'QB34W3I#375%;J838Q343I#:&AJ K*A%383#NM"V(TT#6%3#'X3;3,7+8#BP3A73B#+##A'+:I#Q3#&BB38J ;',3#,#"#L:;34,&;7W3:D(#&B#B%'>,#7,#P&;%#Q(#&#EGFGH#ADA:'J Q:3I#QD#&#EGFGH#ADA:'&II7;7',#Q3;>33,#'X'A&4Q3,7+8#+! &II7;7',#Q3;>33,#;%3#&:@D,3#&,I#&:I3%DI3#C'48B#'X3;3,3 &,I#B7:'XD#&:@D,3#,!" T. ,." #b:3A;4'ADA:7A#47,5#'P3,7,5# 'C# ;%3# 'X3;3,3# 8'73;D LC;34# A',B7I34&Q:3# 3CC'4;# 7,# A',I7;7',# 'P;787S&;7',(# >3 57W3B#Į(ȕJ+,B&;+4&;3I#3B;34#+""#=3X;(#+P',#&A7I#&A;7W&;7', >343# P:3&B3I# ;'# C7,I# ;%&;# ;%3# I3B743I# :&A;',3# C'48&;7', 'C#;%3#'X3;&,3(#;%3#:',3#P&74#'C#;%3#A&4Q',D:#54'+P#A&,#7,7J P4'A3BB#A'+:I#Q3#43&:7S3I#>7;%#%75%#3CC7A73,AD#+B7,5#ADA:7A ;7&;3#&,#YJ%)*J&%&#ADA:7S&;7',#;'#57W3#;%3#B&83#7,;3483I7&;3 &A3;&:B#+6#Kg2#e#Zh3#'4#ZLAM#&B#;%3#'X'A&4Q3,7+8#P43J

!"#$%&&#+",#(#'((#-.-/-.0 123'45#6%7383#934:&5##*;+;;5&4;#<#=3>#?'4@ !"#$%&'! !"#$%&'(&")*"+$,-'$."/*012$(3/$4(')"25*3)$4(6&-3"$%73&8"+*+ ()*

+ O H OH H O O C O+ O – NTf2 R TIPS + O TIPS O 16 R Expansion to Medium Sized Lactone Rings 17 !"# !"#$%&'(#)"#*+, $%&'()*$ H+ 15 •• path a OTIPS OH 8':3A+:3B#;'#43&A;#3CC7A73,;:D#>7;%#7B'AD&,&;3B#-#7,#&#E-FGH OTIPS R OTIPS TIPSNTf O HNTf2 2 C&B%7',#;'#C'48#43I+A3IJ%DI&,;'7,#BA&CC':I#.#E*A%383#G OTIPS HNTf2 O N (10 mol%) KLMH" #O,BP743I#QD#?+I7,RB#B+AA3BB(#>3#%DP';%3B7S3I#;%&; O + O § One of the first intermolecular medium O Ar CH2Cl2, r.t. KT(,MJ&8P%';347A#8':3A+:3B#K,#UVM#8&D#+,I345'#ADA:7S&J O Ar O ring O R O O ;7',#>7;%#I7P':&4'P%7:3B#;'#C'48#83I7+8#&,I#:&453#47,5B" 12 12 13 R !3#3,W7B7',3I#B;4+A;+43#,"(#>%7A%#7,A'4P'4&;3B#&,#&:I3J R R 14 14 TIPSNTf2 18 %DI3#&,I#&,#'X3;&,3#+,7;#;'53;%34#>7;%#&#;>'J&;'8#:7,@J § Acid catalyst pathscreening: b 0(Y 34" #Z,A3#;%3#&:I3%DI3#54'+P#7B#&;;&A@3I#QD#& R OH OTIPS OTIPS – HNTf2 was [2+2]best O ,+A:3'P%7:3(#;%3#43B+:;7,5#,+A:3'P%7:7A#'XD53,#&;'8#7B#3XJ + H+ H P3A;3I#;'#7,7;7&;3#'X3;&,3#'P3,7,5#&,I#53,34&;3#&,';%34 – TfOH / AuOTfOTIPS / AuCl3 / AgNTf2 + F 13 O O ,+A:3'P%7:7A#'XD53,#&;'8#&;#;%3#NJP'B7;7'," O O worked in lower yields O •• •• O TIPS O O – TsOH / MsOH / TFA trace yields O (A) Yudin's design: (1,3)-amphoteric molecules R R – (14e), 62% O TIPS O TIPS NHTf R R 19 21 2 R1 R: R: R R1 n-Bu (14a), 71% n-Bu (14f), 70% 20 3 O § One large limitation [3+2] t-Bu (14b), 60% t-Bu (14g), 87% NH N + C n-Oct (14c), 64% n-Oct (14h), 74% !89:/:;- $e=(0+*Q="$1"68(3*+1 2 1 O – Requires aryl backbone N Ph (14d), 47% Ph(CH2)3 (14i), 51% R2 N TIPSO(CH ) (14j), 71% O HO 2 3 R2 O O '*3)2"VK(3+*-3$K'-&-6-=$6(3$(=+-$Q"$(KK=*"/$&-$&8"$+73&8"2 7 89 OTIPS OTIPS +*+$-,$=(6&-3"+$#*&8$-&8"'$'*3)$+*W"+$R+*0D1SO$G-'$"V(1K="A Ar +&('&*3)$,'-1$&8"$,*N"2A$ +"N"32A$"*)8&2A$ (3/$XM21"1Q"'/ (B) Our desgin: (1,6)-amphoteric molecules MeO O O O O '*3)$(6"&(=+A$&8"$6-''"+K-3/*3)$+"N"32A$3*3"2A$&"32A$(3/$XY2 5 6 12 22 S 1"1Q"'"/$=(6&-3"+$6(3$(==$Q"$,-'1"/$*3$)--/$&-$"V6"=="3& O MeO O O <5% yield 4 [6+2] 7*"=/+O$Z3$&"'1+$-,$&8"$(=@"3"$6-3,*)0'(&*-3$*3$&8"$=(6&-3" R 3 + R !89:/:;* $I,,"6&$-,$&8"$=*3@"' K'-/06&+A$ +"N"32$ &-$ 3*3"21"1Q"'"/$ =(6&-3"+$ #"'"$ -Q2 2 R: R: 1 O § Mechanism is question!! n-Bu (14k), 47% n-Bu (14m), 49% &(*3"/$(+$($+*3)="$)2(=@"3"A$(3/$&"321"1Q"'"/$=(6&-3"$+*2 t-Bu (14l), 52% t-Bu (14n), 53% #(+$-Q&(*3"/$(+$($X[X$1*V&0'"$-,$)\*$*+-1"'+O$Z3$6-3&'(+&A 10 11 Ph (14o), 48% a Sun, J. Angew. Chem. Int. Ed. 2012, 51, 6209.! R R -3=7$&8"$*$*+-1"'$#(+$-Q+"'N"/$,-'$&8"$=(')"2'*3)$=(6&-3" b +*1O $/01213+ #L8P%';347A#8':3A+:3B#7,#47,5#C'48&;7',#43&A;7',B $/01213! #b75%;J838Q343I#:&A;',3#BD,;%3B7B#W7&#ENFGH#ADA:7S&;7', b n+2 O OTIPS O O ;8"$=(6&-3"$,-'1(&*-3$'"(6&*-3$6-0=/$Q"$,0'&8"'$0&*=*W"/ 6%343C'43(#;%3#B;4+A;+43#A'+:I#Q3#A',B7I343I#KT(NMJ&8P%'J ,-#W7&#+,"#!3#%&W3#I',3#B'83#83A%&,7B;7A#B;+I73B"#c'4 24 25 ,-'$*&"'(&*N"$'*3)$"VK(3+*-3+$R%68"1"$YSO$]3/"'$&8"$+&(32 ;347A#&,I#7;#7B#P'BB7Q:3#C'4#7;#;'#+,I345'#ENFGH#ADA:7S&;7', 3X&8P:3(#>3#%&W3#BD,;%3B7S3I#&:@3,3#+"#Kd#e#]M#&,I#B+QJ /('/$6-3/*&*-3+A$&8"$'"(6&*-3$-,$&8"$,*N"21"1Q"'"/$676=*6 >7;%# I7P':&4'P%7:3B# ;'# C'48# 83I7+8J47,5# 8':3A+:3B# ,, f3A;3I#7;#;'#;%3#B;&,I&4I#43&A;7',#A',I7;7',B"#6%3#I3B743I (6"&(=$+30$(3/$(=@73"$,(.$(,,-'/"/$+"N"321"1Q"'"/$=(62 E*A%383#G#K[MH" P4'I+A;#A'+:I#Q3#'QB34W3I(#B+553B;7,5#;%&;#+"#875%;#Q3#& a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n G ;348':3A+:&4#43&A;7',B#C'4#83I7+8J47,5#:&A;',3#BD,;%3B7B( O + "VK(3+*-3$+"U0"36"$6-0=/$Q"$6-3+*/"'"/$($,-'1(=$_`aJaJb C'48#&,53#'C#375%;J838Q343I#:&A;',3B#,#"^#_,I34#87:I ;%3#43&A;7',#B+CC34B#C4'8#&f'4#:787;&;7',(#7"3"(#&,#&4D: 23 TIPSO 676=*W(&*-3$'"(6&*-3O$Z3$K'*36*K="A$&8"$*&"'(&*N"$'*3)2"VK(32 A',I7;7',B(# ;%3# 43&A;7',# 3X%7Q7;B# 5''I# C+,A;7',&:# 54'+P :7,@34#Q3;>33,#;%3#'X3;&,3#&,I#&:I3%DI3#8'73;73B#7,#;%3 13 +*-3$K'-&-6-=$6-0=/$Q"$,0'&8"'$(KK=*"/$,-'$1-'"$&*1"+$&- A'8P&;7Q7:7;D"# *;347A&::D# %7,I343I# &:@D,3B# &43# &:B'# B+7;J &8P%';347A# 8':3A+:3B# 7B# 43`+743I# ;'# &A%73W3# 43&B',&Q:3 !89:/:;3 $f"+*)3$-,$($3"#$'*3)$"VK(3+*-3$+&'(&")7 (++"1Q="$(==$+*W"+$-,$=(6&-3"+$,'-1$+1(==$676=*6$(6"&(=+O &Q:3#B+QB;4&;3B"#]3;34'ADA:3JC+B3I#:&A;',3B#K,#2/4M#A&, 3CC7A73,AD"#c'4#3X&8P:3(#;%3#+B3#'C#&,#&:@D:#:7,@34#K3"5"( Downloaded by: IP-Proxy University of Texas at Austin, Austin Libraries. Copyrighted material. &:B'#Q3#C'483I#B+AA3BBC+::D#K*A%383#-M" ++(#*A%383#VM#43B+:;B#7,#,'#I3B743I#:&A;',3#C'48&;7',#+,J ;8"$K'-K-+"/$1"68(3*+1$*+$/"K*6&"/$*3$%68"1"$c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aJb$676=-(//*&*-3$#*&8$-V-6('Q"2 ;'#C'48#'X'A&4Q3,7+8#,5(#C'::'>3I#QD#&;;&A@#'C#;%3#3:3AJ 8*)8$ ",,*6*"367$ R+*.D/SO$ P$ /*N"'+"$ +"&$ -,$ ,036&*-3(= 3*01$+4$&-$,0'3*+8$&8"$@"7$-V"&"3*01$*3&"'1"/*(&"$()O$G*2 ;4',J47A%#&:@D,3#;'#57W3#@3;3,7+8#7,;3483I7&;3#,6"#*+QB3J L787,5#&;#C+4;%34#78P4'W7,5#;%3#3CC7A73,AD#&,I#BA'P3#'C )'-0K+$('"$6-1K(&*Q="$#*&8$&8"$'"(6&*-3$6-3/*&*-3+O$T-12 3(==7A$ "="6&'-676=*6$ '*3)$ -K"3*3)$ -,$ &8"$ -V"&"3*01$ '*3) `+3,;#ADA:7S&;7',#Q3;>33,#;%3#438';3#%DI4'XD:#54'+P#&,I 83I7+8J47,5#:&A;',3#BD,;%3B7B(#P&4;7A+:&4:D#;'#&II43BB#;%3 K('*3)$#*&8$&8"$K'"N*-0+$+&'(&")7A$&8"$K'"+"3&$1"&8-/$/-"+ /"=*N"'+$&8"$-Q+"'N"/$'*3)2"VK(3+*-3$K'-/06&$+*O$d"$K'-2 ;%3#@3;3,7+8#8'73;D#C'48B#ADA:7A#B7:D:#@3;3,3#&A3;&:#,7" &Q'W3# B+QB;4&;3# :787;&;7',(# >3#%&W3#43A3,;:D#43P'4;3I#& 3-&$'"U0*'"$(3$('7=$)'-0K$,0+"/$*3$&8"$=(6&-3"$+@"="&-3ADownloaded by: IP-Proxy University of Texas at Austin, Austin Libraries. Copyrighted material. K-+"$&8(&$&8"$'-="$-,$(//*&*N"$JALAM26-==*/*3"$*+$&-$+&(Q*=*W" =3X;(#&A;7W&;7',#'C#,7#QD#P4';',&;7',#'C#;%3#3;%34#8'73;D A',A3P;+&::D# I7CC343,;# B;4&;35D# A%&4&A;347S3I# QD# &,# +,J #8*68$ +*)3*,*6(3&=7$ "VK(3/+$ &8"$ +6-K"O$ .-'"-N"'A$ &8" &8"$-V-6('Q"3*01$+4$Q7$'"N"'+*Q=7$,-'1*3)$&8"$K7'*/*3*2 ;475534B#;%3#a௅Z#Q',I#A:3&W&53#;'#C'48#:&A;',3#,-(#>%7A% P43A3I3,;3I# 47,5# 3XP&,B7',# 'C# Q7ADA:7A# 'X3;3,7+8# +# 7B#B7:D:&;3I#;'#I3:7W34#;%3#'QB34W3I#375%;J838Q343I#:&AJ K*A%383#NM"V(TT#6%3#'X3;3,7+8#BP3A73B#+##A'+:I#Q3#&BB38J ;',3#,#"#L:;34,&;7W3:D(#&B#B%'>,#7,#P&;%#Q(#&#EGFGH#ADA:'J Q:3I#QD#&#EGFGH#ADA:'&II7;7',#Q3;>33,#'X'A&4Q3,7+8#+! 9:"-')$;8*"1"$<"'=()$$%&0&&)('&$>$!"#$?-'@ !"#$%&&$+),-A$'(A$BCBDBCE &II7;7',#Q3;>33,#;%3#&:@D,3#&,I#&:I3%DI3#C'48B#'X3;3,3 &,I#B7:'XD#&:@D,3#,!" T. ,." #b:3A;4'ADA:7A#47,5#'P3,7,5# 'C# ;%3# 'X3;3,3# 8'73;D LC;34# A',B7I34&Q:3# 3CC'4;# 7,# A',I7;7',# 'P;787S&;7',(# >3 57W3B#Į(ȕJ+,B&;+4&;3I#3B;34#+""#=3X;(#+P',#&A7I#&A;7W&;7', >343# P:3&B3I# ;'# C7,I# ;%&;# ;%3# I3B743I# :&A;',3# C'48&;7', 'C#;%3#'X3;&,3(#;%3#:',3#P&74#'C#;%3#A&4Q',D:#54'+P#A&,#7,7J P4'A3BB#A'+:I#Q3#43&:7S3I#>7;%#%75%#3CC7A73,AD#+B7,5#ADA:7A ;7&;3#&,#YJ%)*J&%&#ADA:7S&;7',#;'#57W3#;%3#B&83#7,;3483I7&;3 &A3;&:B#+6#Kg2#e#Zh3#'4#ZLAM#&B#;%3#'X'A&4Q3,7+8#P43J

!"#$%&&#+",#(#'((#-.-/-.0 123'45#6%7383#934:&5##*;+;;5&4;#<#=3>#?'4@ !"#$%&'! !"#$%&'(&")*"+$,-'$."/*012$(3/$4(')"25*3)$4(6&-3"$%73&8"+*+ ()*

+ O H OH H O O C O+ O – NTf2 R TIPS + O TIPS O 16 R 17 H+ 15 •• path a OTIPS OH R OTIPS TIPSNTf HNTf2 2 O O O O O O 12 R R 14 TIPSNTf2 18 path b R OH [2+2] O + H+ H OTIPS + 13 O O

O •• O TIPS ••O R R – O TIPS O TIPS NHTf 19 21 2 R 20 !89:/:;- $e=(0+*Q="$1"68(3*+1

O O '*3)2"VK(3+*-3$K'-&-6-=$6(3$(=+-$Q"$(KK=*"/$&-$&8"$+73&8"2 +*+$-,$=(6&-3"+$#*&8$-&8"'$'*3)$+*W"+$R+*0D1SO$G-'$"V(1K="A Ar +&('&*3)$,'-1$&8"$,*N"2A$ +"N"32A$"*)8&2A$ (3/$XM21"1Q"'/ O O '*3)$(6"&(=+A$&8"$6-''"+K-3/*3)$+"N"32A$3*3"2A$&"32A$(3/$XY2 12 22 1"1Q"'"/$=(6&-3"+$6(3$(==$Q"$,-'1"/$*3$)--/$&-$"V6"=="3& Improved Access to Large Lactones <5% yield 7*"=/+O$Z3$&"'1+$-,$&8"$(=@"3"$6-3,*)0'(&*-3$*3$&8"$=(6&-3" !89:/:;* $I,,"6&$-,$&8"$=*3@"' K'-/06&+A$ +"N"32$ &-$ 3*3"21"1Q"'"/$ =(6&-3"+$ #"'"$ -Q2 &(*3"/$(+$($+*3)="$)2(=@"3"A$(3/$&"321"1Q"'"/$=(6&-3"$+*2 #(+$-Q&(*3"/$(+$($X[X$1*V&0'"$-,$)\*$*+-1"'+O$Z3$6-3&'(+&A a R R -3=7$&8"$*$*+-1"'$#(+$-Q+"'N"/$,-'$&8"$=(')"2'*3)$=(6&-3" § Needed to fix aryl linker substrate limitation b +*1O b n+2 O – Form various sizes of rings both large and small OTIPS O O ;8"$=(6&-3"$,-'1(&*-3$'"(6&*-3$6-0=/$Q"$,0'&8"'$0&*=*W"/ 24 25 ,-'$*&"'(&*N"$'*3)$"VK(3+*-3+$R%68"1"$YSO$]3/"'$&8"$+&(32 – Does not need any forced configuration /('/$6-3/*&*-3+A$&8"$'"(6&*-3$-,$&8"$,*N"21"1Q"'"/$676=*6 (6"&(=$+30$(3/$(=@73"$,(.$(,,-'/"/$+"N"321"1Q"'"/$=(62 a [2+2] &-3"$+*0A$#8*68$#(+$"(+*=7$6-3N"'&"/$&-$(6"&(=$+4$(,&"' !"# !"#$%&'(#)"#*+, $%&'()*$ 87/'-)"3(&*-3$(3/$'"/06&*N"$(67=(&*-3O$%0Q^"6&*-3$-,$(6"2 R &(=$ +4$ &-$ ($ +"6-3/$ '*3)$ "VK(3+*-3$ #*&8$ (=@73"$ ,(5$,0'2

R2 3*+8"/$ 3*3"21"1Q"'"/$ =(6&-3"$ +*6O$ ;8"$ *&"'(&*N"$ '*3)2 R1 BF ⋅OEt (2.0 equiv) n + 3 2 R2 O 2,4,6-collidine (1.0 equiv) "VK(3+*-3$+"U0"36"$6-0=/$Q"$6-3+*/"'"/$($,-'1(=$_`aJaJb n + R1 n+2 23 TIPSO 676=*W(&*-3$'"(6&*-3O$Z3$K'*36*K="A$&8"$*&"'(&*N"$'*3)2"VK(32 13 CH2Cl2 (0.1 M) O O LG TIPSO O +*-3$K'-&-6-=$6-0=/$Q"$,0'&8"'$(KK=*"/$,-'$1-'"$&*1"+$&- 26 13 25 !89:/:;3 $f"+*)3$-,$($3"#$'*3)$"VK(3+*-3$+&'(&")7 (++"1Q="$(==$+*W"+$-,$=(6&-3"+$,'-1$+1(==$676=*6$(6"&(=+O ;8"$K'-K-+"/$1"68(3*+1$*+$/"K*6&"/$*3$%68"1"$cO$;8"$'"2 Downloaded by: IP-Proxy University of Texas at Austin, Austin Libraries. Copyrighted material. 60'+-'$ (3/$ 4"#*+$ (6*/$ FGB>HI&J$(+$&8"$K'-1-&"'$#*&8 (6&*-3$Q")*3+$#*&8$FG 2K'-1-&"/$"=*1*3(&*-3$-,$(=@-V*/" 1 B R2 (25c), 89% n-Bu R : OTBS (25f), 75% Me JALAM26-==*/*3"$(+$&8"$(//*&*N"O$P+$+8-#3$*3$%68"1"$EA$( ,'-1$(6"&(=$ +3$&-$,-'1$-V-6('Q"3*01$+4O$Z3$ &8"$1"(32 8 (25d), 58% 8 OBz (25g), 56% 8 '(3)"$-,$"*)8&21"1Q"'"/$=(6&-3"+$#*&8$/*,,"'"3&$+0Q+&*&028 Oct #8*="A$+*=-V7$(=@73"$,($*+$(6&*N(&"/$&-$,-'1$73-=(&"$+7A O R1 O N3 (25h), 52% n-Bu n-Bu O TIPSO (25e), 62% O O "3&+$(&$/*,,"'"3&$K-+*&*-3+$6(3$Q"$-Q&(*3"/$#*&8$)--/$&-O OCH2CH=CH2 (25i), 65% #8*68$&8"3$03/"')-"+$_JaJb$676=-(//*&*-3$#*&8$-V-6('Q"2 R2: n-Bu (25a), 91% O 8*)8$ ",,*6*"367$O R+*.D/SO$ P$ /*N"'+"$ +"&$ -,$ ,036&*-3(= CH2CO CH (25j), 63% 3*01$+4$&-$,0'3*+8$&8"$@"7$-V"&"3*01$*3&"'1"/*(&"$()O$G*2 25k (77%) 25l (72%) Ph (25b), 64% )'-0K+$('"$6-1K(&*Q="$#*&8$&8"$'"(6&*-3$6-3/*&*-3+O$T-12 3(==7A$ "="6&'-676=*6$ '*3)$ -K"3*3)$ -,$ &8"$ -V"&"3*01$ '*3) K('*3)$#*&8$&8"$K'"N*-0+$+&'(&")7A$&8"$K'"+"3&$1"&8-/$/-"+ n-Bu /"=*N"'+$&8"$-Q+"'N"/$'*3)2"VK(3+*-3$K'-/06&$+*O$d"$K'-2 3-&$'"U0*'"$(3$('7=$)'-0K$,0+"/$*3$&8"$=(6&-3"$+@"="&-3An-Bu n-Bu K-+"$&8(&$&8"$'-="$-,$(//*&*N"$JALAM26-==*/*3"$*+$&-$+&(Q*=*W" n-Bu O n-Bu n-Bu #8*68$ +*)3*,*6(3&=7$O "VK(3/+$ &8"$ +6-K"O$ .-'"-N"'A$ &8" 7 &8"$-V-6('Q"3*01$+4$Q7$'"N"'+*Q=7$,-'1*3)$&8"$K7'*/*3*2 O Me 7 9 10 18 1 O O O 8 R O O O O Me O O R1 = H (25n), 93% b 25q (86%) 25rb (88%) 9:"-')$;8*"1"$<"'=()$$%&0&&)('&$>$!"#$?-'@25sb (87%) !"#$%&&$+),-A$'(A$BCBDBCE 25m (91%) R1 = Ph (25o), 70% Me Me Z/E = 1:1 25p (70%)

$1A7:7B0 #*B'E3#'L#;%3#47,5S3PE&,D7',#43&B;7',"&#&#g2#i#je3#+,:3DD#';%34>7D3#,';3A"#K#g2#i#jCB" Sun, J. JACS 2013, 135, 4680.!

n-Bu +8#&AA+B;#+."#C#D787:&4#&AA+B;#>&D#E43F7'+D:G#43E'4;3A [5+2+2] IJ OTIPS &,A#B%&4&B;347H3A" #C:;%'+5%#G,':&;3#+/#7D#E4'E'D3A#;'#K3 O OAc 9 ;%3#&B;7F3#DE3B73D#L'4#;%3#MJNJO#BGB:'&AA7;7',(#>3#B'+:A O O 26n ,';# 4+:3# '+;# ;%3# E'DD7K7:7;G# 'L# BGB:'&AA7;7',# K3;>33, 25t (65%) D7:'PG#&:@G,3#,-#&,A#'P'B&4K3,7+8#+0#L'::'>3A#KG#47,5 n-Bu 3PE&,D7',QA3D7:G:&;7'," first second OTIPS expansion expansion R,#B',B:+D7',(#>3#%&F3#D+88&47H3A#%343#'+4#43B3,;#A3F3:S TIPSO TIPSO 'E83,;#'L#;>'#,3>#D;4&;3573D#L'4#83A7+8S#&,A#:&453S47,5 13j 13a :&B;',3#DG,;%3D7D#38E:'G7,5#D7:'PG#&:@G,3D"#T,&K:3A#KG ;%3#A3D75,#'L#&#;GE3#'L#,3>#UI(VWS&8E%';347B#8':3B+:3D n-Bu 1) Pd/C, H2 n-Bu 7 7 B',;&7,7,5#'P3;&,3#&,A#&:A3%GA3#L+,B;7',&:7;73D(#;%3#L74D; 2) DIBAL-H; O O D;4&;35G# 43E43D3,;D# ',3# 'L# ;%3# L74D;# 7,;348':3B+:&4# 43&BS O Ac2O, Py OAc 25n (93%) 68% for 2 steps 27 (dr = 5:1) ;7',D#L'4#375%;S838K343AS47,5#:&B;',3#L'48&;7',"#X'>3FS 34(#;%3#43&B;7',#7D#D'83>%&;#:787;3A#KG#;%3#43Y+74383,;#'L $1A7:7B. #R;34&;7F3#47,5#3PE&,D7',D &,#&4G:#:7,@34#7,#;%3#&8E%';347B#8':3B+:3D"#=3P;(#>7;%#&, +,E43B3A3,;3A#L+D3A#'P3;3,7+8#47,5S3PE&,D7',#D;4&;35G( R1 >3#A3F3:'E3A#&#D3B',A#D;4&;35G#;%&;#,';#',:G#3LL7B73,;:G 1 BF3 1 R R Me L'48D# 83A7+8S# &,A# :&453S47,5# :&B;',3D(# K+;# &:D'# 'F34S 2,4,6-collidine B'83D#;%3#:787;&;7',D#7,#;%3#L74D;#D;4&;35G"#Z';%#D;4&;3573D O N O OR O &43#7,;348':3B+:&4#43&B;7',D#>7;%'+;#+D7,5#%75%#A7:+;7',#'4 26 Me Me 27 D:'>#&AA7;7',#;%&;#7D#B',D;&,;:G#38E:'G3A#7,#B',F3,;7',&: 28 &EE4'&B%3D#;'#3,%&,B3#43&B;7',#3LL7B73,BG"#!3#&,;7B7E&;3 ;%&;#;%3D3#43&B;7',D#>7::#L7,A#&EE:7B&;7',D#7,#'45&,7B#DG,S RO BF3 1 1 ;%3D7D" R2 R2 R R

2 R Downloaded by: IP-Proxy University of Texas at Austin, Austin Libraries. Copyrighted material. O O (123456789:73; 2 RO BF R OTIPS 2 O O [7,&,B7&:#D+EE'4;#>&D#E4'F7A3A#KG#X\]*6#U^C2II*_.IW#&,A#;%3 13 TIPSF 29 25 O =&;7',&:#=&;+4&:#*B73,B3#['+,A&;7',#'L#_%7,&#UJII.J.J`W" LA 30

$1A7:7B/ #c4'E'D3A#83B%&,7D8 <7=7>7317? UIW U&W a'+DD3&+(#2")*%&+,-%.+/##,//@()(0(#J000"#UKW *%77,&(#R") 1-%23)4%53#+""0()067(#J-b"#UBW c&43,;G(#C"d#e'43&+(#f"d# =73:(#2"d#_&8E&5,3(#)"Se")1-%23)4%53#+",!()008(#caId#&,A# 43L343,B3D#B7;3A#;%3437," UJW R::+87,&;7(#2"d#e&,A':7,7(#g")9::3)1-%23)4%;3#,/.,()0<(#bh"

!"#$%&&#+",-(#'((#-.-/-.0 123'45#6%7383#934:&5##*;+;;5&4;#<#=3>#?'4@ Journal of the American Chemical Society Communication

Table 3. Synthesis of Medium and Large Ring Lactones

good yield under our standard conditions. In contrast, from cyclic orthoester 11, we anticipated the formation of a lactone product if the OMe could serve as the leaving group and thus the Improved Access to Largering-fused Lactones oxetenium G could be formed (eq 8). As expected, the desired lactone 12 was obtained in 82% yield. Finally, subjecting linear unsaturated ester 13 to our standard conditions did not § Utilized linear & cyclic control experimentslead toto theunravel formation the of mechanism lactone 3a (eq 9), which ruled out the possible mechanism involving an initial acetal olefination followed by lactonization. Journal of the American Chemical Society § Key intermediate is oxeteniumCommunication species On the basis of the above observations, we have proposed a plausible mechanism (Scheme 3). The reaction begins with the Table 3. Synthesis of Medium and Large Ring Lactones Scheme 3. Proposed Mechanism

• [2+2] may be through stepwise ketene species

BF3-promoted formation of oxocarbenium H from acetal 1. The aIsolated yield, Z/E > 20:1. bRun at room temperature. cZ/E = 1:1. resulting trifluoro(methoxy)borate next activates the siloxy d e • Collidine is proposed to stabilize the highly unstable E/Z > 20:1. No desired lactone formation; 1w was recovered. alkyne to form ynolate J. The formation of TIPSF can be good yield under our standard conditions. In contrast, from observed.oxocarbenium Subsequent [2+2] cycloaddition between the two Schemecyclic 2. orthoesterIterative Ring11,Expansions we anticipated the formation of a lactone highly active species (H and J) forms oxetenium intermediate L, product if the OMe could serve as the leaving group and thus the presumably by a stepwise mechanism via ketene K. Finally, ring-fused oxetenium G could be formed (eq 8). As expected, the electrocyclic ring-opening of L delivers the observed ring- desired lactone 12 was obtained in 82% yield. Finally, subjecting expansion product 3. While the exactSun, role J. JACS played 2013 by, 135 2,4,6-, 4680.! linear unsaturated ester 13 to our standard conditions did not collidine is still not clear, we propose that it can reversibly form lead to the formation of lactone 3a (eq 9), which ruled out the the pyridinium I with the highly unstable oxocarbenium H. Thus, possible mechanism involving an initial acetal olefination adduct I can be considered a reservoir of H to prevent its followed by lactonization. decomposition before reacting with ynolate J. We were able to On the basis of the above observations, we have proposed a observe a peak at δ ∼6.0 ppm by in situ 1HNMR, which is plausible mechanism (Scheme 3). The reaction begins with the characteristic for this type of pyridinium species.14 On the basis of the report by Fujioka and Kita, pyridine can also form a similar Scheme 3. Proposed Mechanism adduct, but this adduct is much less reactive than that from 2,4,6- that cyclic acetal 7 and acyclic mixed acetal 9 would give linear collidine.14 This result is consistent with our observation that the ester products if the reactions proceed via the hypothesized use of pyridine as additive resulted in no desired lactone oxetenium intermediates E and F, respectively (eqs 6 and 7). formation (Table 1, entry 16). We believe that the superiority of Indeed, the expected acyclic esters 8 and 10 were obtained in 2,4,6-collidine is attributed to the excellent stabilization of the

4682 dx.doi.org/10.1021/ja400883q | J. Am. Chem. Soc. 2013, 135, 4680−4683

BF3-promoted formation of oxocarbenium H from acetal 1. The aIsolated yield, Z/E > 20:1. bRun at room temperature. cZ/E = 1:1. resulting trifluoro(methoxy)borate next activates the siloxy d e E/Z > 20:1. No desired lactone formation; 1w was recovered. alkyne to form ynolate J. The formation of TIPSF can be observed. Subsequent [2+2] cycloaddition between the two Scheme 2. Iterative Ring Expansions highly active species (H and J) forms oxetenium intermediate L, presumably by a stepwise mechanism via ketene K. Finally, electrocyclic ring-opening of L delivers the observed ring- expansion product 3. While the exact role played by 2,4,6- collidine is still not clear, we propose that it can reversibly form the pyridinium I with the highly unstable oxocarbenium H. Thus, adduct I can be considered a reservoir of H to prevent its decomposition before reacting with ynolate J. We were able to observe a peak at δ ∼6.0 ppm by in situ 1HNMR, which is characteristic for this type of pyridinium species.14 On the basis of the report by Fujioka and Kita, pyridine can also form a similar adduct, but this adduct is much less reactive than that from 2,4,6- that cyclic acetal 7 and acyclic mixed acetal 9 would give linear collidine.14 This result is consistent with our observation that the ester products if the reactions proceed via the hypothesized use of pyridine as additive resulted in no desired lactone oxetenium intermediates E and F, respectively (eqs 6 and 7). formation (Table 1, entry 16). We believe that the superiority of Indeed, the expected acyclic esters 8 and 10 were obtained in 2,4,6-collidine is attributed to the excellent stabilization of the

4682 dx.doi.org/10.1021/ja400883q | J. Am. Chem. Soc. 2013, 135, 4680−4683 Catalytic Asymmetric Ring Opening

§ 3-Substituted oxetanes are prochiral! – Ring opening can lead to chiral products (desymmetrization) – Form chiral, high substituted 3 carbon building blocks

§ Challenges: – Alcohol product is competing nucleophile – limiting strong nucleophiles for opening or internal nucleophiles View Article Online – Chiral lewis acid coordination is remote to the generated chiral center Organic & Biomolecular Chemistry Perspective

Scheme 1 Catalytic asymmetric nucleophilic openings of 3-substituted Scheme 2 Asymmetric opening with a lithium reagent by assistance of oxetanes. chiral ligands and boron reagents.

efficient chiral induction is challenging. To date, there are few examples of such catalyst-controlled stereoselective nucleo- philic openings of 3-substituted oxetanes.3

2. Catalytic asymmetric nucleophilic openings of oxetanes

In 1996, Tomioka and co-workers reported their pioneering study of enantioselective ring-opening reactions of 3-substi- tuted oxetanes.4 A chiral boron reagent generated from phenyl-

lithium, BF3, and chiral ether 4 reacted with 3-phenyloxetane 3 to form the alcohol 5 in excellent yield but with low enantio- selectivity (47% ee, Scheme 2). Although the reaction required a super-stoichiometric amount of the chiral boron reagent and proceeded with moderate stereoselectivity, it represented the first intermolecular asymmetric nucleophilic desymmetriza- tion of 3-substituted oxetanes.

In 2009, Loy and Jacobsen reported an intramolecular Scheme 3 (Salen)Co(III)-catalyzed intramolecular desymmetrization. desymmetrization of 3-substituted oxetanes catalyzed by 5 (salen)Co(III) complexes (Scheme 3). With either the mono- meric or the oligomeric form (i.e., 8 or 9) of the Lewis acid stereocenters could also be generated with up to 99% ee. The catalyst, oxetanes 6 underwent intramolecular opening to form oligomeric complex 9 proved superior to the monomeric form a range of tetrahydrofuran products 7 with both excellent 8 and the catalyst loading could be as low as 0.01 mol%. The Published on 12 June 2014. Downloaded by University of Texas Libraries 18/09/2014 20:37:53. efficiency and remarkable enantioselectivity. Quaternary better performance of the oligomeric catalyst, particularly in the case of tetrahydropyran synthesis, might be attributed to its capability of extending the chiral backbone for remote Jianwei Sun graduated with B.S. chiral induction. However, although the process was efficient and M.S. degrees in Chemistry for the synthesis of a range of highly enantioenriched tetra- from Nanjing University in 2001 hydrofurans and a tetrahydropyran, it could not be extended and 2004, respectively. In 2008, with similar efficiency to the formation of seven-membered he obtained his Ph.D. degree in ring oxepanes or pyrrolidines using an internal nitrogen Organic Chemistry from the Uni- nucleophile. The attempt to use intermolecular nucleophiles versity of Chicago, working with also failed. Nevertheless, it is the first demonstration of a Professor Sergey A. Kozmin. He truly catalytic asymmetric nucleophilic opening of 3-substi- then worked as a postdoctoral tuted oxetanes with excellent stereocontrol. fellow with Professor Gregory Recently, our lab has reported a series of studies of this C. Fu at the Massachusetts Insti- type employing chiral Brønsted acid catalysis.6–8 While the tute of Technology. In 2010, he basicity of oxetanes is relatively higher than epoxides and 9 Jianwei Sun became an Assistant Professor of normal ethers, presently widely used chiral Brønsted acids, Chemistry at the Hong Kong Uni- such as chiral phosphoric acids, are relatively weak in acidity. versity of Science and Technology. He is a recipient of the Asian This means either strong or internal nucleophiles are necess- Core Program Lectureship Award, Hong Kong Research Grants ary. We initially focused on the evaluation of internal nucleo- Council Early Career Award, and Thieme Chemistry Journal philes. By mixing the oxetane-tethered aryl aldehydes 10,10 Award. aryl amines 11, indoles 12, and a catalytic amount of chiral

This journal is © The Royal Society of Chemistry 2014 Org. Biomol. Chem.,2014,12,6028–6032 | 6029 Minireviews M. Rogers-Evans et al.

4.3. Ring-Expansion Reactions of Oxetanes

As part of their studies of transition-metal-catalyzed enantioselective reactions, Noyori and co-workers document- ed the copper-catalyzed asymmetric ring expansion of ox- etanes to afford tetrahydrofurans.[66] They found that the Scheme 16. Selective opening of 2-phenyloxetane with aryl borates. insertion of a salicylaldimine-chelated copper carbenoid into the C O bond can effect the formation of enantiomerically À enriched substituted tetrahydrofurans. Inspired by this work, 4.2. Intramolecular Ring-Opening Reactions Katsuki and co-workers were interested in further improving

the reaction. The use of C2-symmetric bipyridine ligands for Bach et al. examined the intramolecular opening of 2,3,3- copper(I) was advantageous for the generation of enantioen- trisubstituted oxetanes to give ring-expanded ethers, thioeth- riched products (Scheme 18).[67] Treatment of racemic 2- ers, and carbonates.[63] Loy and Jacobsen documented the phenyloxetane (18) with tert-butyl diazoacetate and a mixture Catalytic Asymmetric Ringenantioselective Opening Lewis acid mediated intramolecular ring of 1 mol% of CuOTf and bipyridine ligand 100 afforded the opening of oxetanes.[64] Achiral 3,3-disubstituted oxetanes bearing a remote hydroxy group were treated at ambient § 1996: Tomioka first intermolecular nucleophilictemperature desymmetrization either neat or at high concentrations (for example, 6m in CH3CN) with catalytic amounts of cobalt(III) Ph salen complexes. The ring-opening reactions afforded a Ph Ph variety of tetrahydrofurans 95 in high yield and enantiose- OMe lectivity (Scheme 17). In accordance with epoxide-opening O O MeO (2.1 equiv.) reactions,[65] dimeric complexes 97 (n = 1) were found to be + PhLi HO Ph BF -OBu (1.5 equiv.) 92% yield, 47% ee 3 2 Ph Ph Et2O, -78 °C Tomoika, K. Tetrahedron: Asymmetry 1996, 7, 2483. Tomioka, K. Tetrahderon 1997, 53, 10699. ! Scheme 18. Asymmetric ring expansion of oxetanes according to Katsuki and co-workers.[67] TBS =tert-butyldimethylsilyl. § 2009: Loy and Jacobsen accomplish intramolecular approach – Oligomeric catalyst extends chiral backbone for remote induction products 98 and 99 in a diastereomeric ratio of 59:41 in favor [67a,b] – Failed for intermolecular reactions of the trans isomer. The conversion of 2-alkynyloxetane into the corresponding tetrahydrofuran proved to be superior, as the product was formed in 88% yield (as a cis/trans mixture). This transformation was applied in the total syn- thesis of the trans-Whisky lactone and the formal total OH syntheses of ( )-avenaciolide and ( )-isoavenaciolide.[67c–e] R O À À R In 2001, Lo and Fu described the use of a copper(I) 8 or 9 (0.01 mol%) bis(azaferrocene) complex in an oxetane ring-expansion MeCN, 23 C [68] OH ° O reaction. By using enantiomerically highly enriched 2- substituted oxetanes and a diazoacetate with 1 mol% of a up to 98% yield copper(I) catalyst, the ring-expanded products could be up to 99% ee obtained in good yield as well as high diastereomeric and enantiomeric ratios (Scheme 19). Sterically demanding car- Scheme 17. Cobalt(III)-catalyzed enantioselective intramolecular open- Jacobsen, E. JACS 2009, 131, 2786.! ing of oxetanes according to Loy and Jacobsen.[64] OTf =trifluoro- methane sulfonate.

superior to the monomeric cobalt(III) salen counterpart (96). Thus, the use of catalysts of type 97 allowed catalyst loadings as little as 0.01 mol% to be used, with the rearranged products formed in up to 98% yield and 99% ee. The enhanced reactivity of catalyst 97 (n = 1) can be explained by cooperative bimetallic mechanisms involving simultaneous activation of the nucleophile as well as Lewis acid activation of the oxetane.

Scheme 19. Ring expansion of oxetanes according to Lo and Fu.[68] Cy = cyclohexyl.

9060 www.angewandte.org  2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Angew. Chem. Int. Ed. 2010, 49, 9052 – 9067 Angewandte. Communications

Further screening of different directing groups showed Table 1: Substrate scope of the reaction with regard to indoles. that the use of a relatively rigid oxetane[12] group can promote the three-component reaction, which proceeded smoothly to form polycyclic product 5 not only with good efficiency, but also with remarkable diastereoselectivity and enantioselec- tivity (99% ee, d.r. > 99:1, Scheme 2, see the Supporting Information for optimization of reaction conditions).[13] It is worth noting that in addition to the desired aza-Diels–Alder Entry R t [h] Product Yield[a] d.r.[b] ee reaction, the process also involves an oxetane desymmetriza- [14] 1 H 12 5 85% >99/1 98% tion by the intramolecular nitrogen nucleophile. It is also 2 5-Me 12 6 89% >99/1 92% noteworthy that the reaction generates four new bonds (two 3 6-Me 12 7 96% >99/1 93% C C and two C N bonds) and four new stereogenic centers in 4 7-Me 12 8 92% >99/1 94% À À one pot from three achiral compounds. 5 5-Br 36 9 76% 90/10 86% Moreover, the polycyclic alkaloid-type product contains 6 6-Br 24 10 63% >99/1 96% Catalytic Asymmetricindoline, tetrahydroquinoline, Ring Opening and tetrahydroisoquinoline 7 6-F 24 11 74% >99/1 94% 8 5-OMe 12 12 82% >99/1 96% moieties, all of which are key elements found in numerous 9 5-OH 12 13 97% >99/1 86% biologically active natural products and synthetic pharma- 10 6-OBn 12 14 78% >99/1 98% § Basicity of oxetanes is ceuticals.higher than[2] To epoxides the best of and our knowledge,ethers there have been no 11 6-COView2Me Article 24Online 15 74% 90/10 74% – Activation by chiralreports phosphoric on such acids an efficient (relatively assembly weak of theseacidity) three requires important strong/internal [a] Yields of isolated purified products. [b] Ratio of the major diastereo- Perspective Organic & Biomolecular Chemistry nucleophiles structural units in one operation. mer to the total of all other diastereomers. Next, we examined the substrate scope of this three- component polycyclization reaction. Differently substituted quinoline moiety can also be generated (25), albeit with moderate enantiomeric excess. It is worth noting that the success of generating quarternary chiral centers by desymmetriza- tion of 3-substituted oxetanes was limited.[14] We also conducted some control experi- ments to probe the reaction mechanism. The reaction of N-methyl-6-bromoindole resulted in the formation of a mixture of products, and the desired N-methylated analogue of prod- uct 10 was obtained in less than 20% yield (compare with entry 6 in Table 1), thus Scheme 2. Oxetane as the crucial directing group. suggesting that the indole N H moiety is À presumably involved in hydrogen bonding. In Scheme§ 4GeneratesChiral Brønsted 2 new acid C-C catalyzed and oxetane 2 new desymmetrization C-N bonds with nitrogen nucleophiles. addition, we found that the reaction exhibits – Chiral product fromindoles 3 achiral participated compounds smoothly in the reaction to form a range a small positive nonlinear effect (NLE, Figure 1). In addition, of polycyclic alkaloid-type molecules with excellent efficiency the reaction with racemic catalyst A proceeded significantly – First example with nitrogen nucleophilesffi ! phosphoric acid A, we were able toand observe stereoselectivity the e cient (Table for- 1). Electron-donating (entries 2– slower that that with enantiopure catalyst under otherwise 6 mation of polycyclic adducts 13 (Scheme4 and 4). 8–10)After and simple electron-withdrawing puri- groupsSun, J. (entries Angew. Chem. 5–7 and Int. Ed. 2013identical, 52, 2027. conditions.! Since we did not observe a solubility fication by filtration or centrifugation,11) all led these to comparably heterocyclic good results. The method is also difference between enantiopure and racemic catalyst A in products were isolated in good tocompatible excellent enantio- with a and diverse dia- setScheme of functional 5 Asymmetric groups, syntheses such as of substituteddiethyl tetrahydroisoquino- ether,[15] the observed NLE may result from the lines and their analogues. stereopurity. Notably, the reactionhalides generated (entries four new 5–7), bonds ethers (entries 8 and 10), esters (two C–C and two C–N bonds) and(entry four 11), new and stereogenic the free hydroxy group (entry 9). It is note- centers in one pot from three achiralworthy compounds. that the Regarding product purification is very easy. All the the reaction mechanism, we proposedstarting the initial materials formation as well of as the catalyst are soluble in diethyl the internal amine nucleophile by sequential imine formation ether, whereas the polycyclic products have low solubility and and indole addition, followed by oxetane opening. Small posi- therefore precipitate out during the reaction. Thus, simple tive non-linear effects were also observed, presumably filtration (or centrifugation) followed by washing typically suggesting a more complicated mechanism. gave pure products. In addition to indoles as the reaction partner, we also The reaction scope in terms of the oxetane-tethered employed Hantzsch ester 15 for the above reaction.7 With a aldehydes and arylamines is shown in Scheme 4. Different similar catalytic system, the reactions of oxetanes 10 and aryl aldehydes could smoothly react to form the desired amines 14 formed a range of tetrahydroisoquinolines 16 with polycyclic products (16–21) efficiently. The reaction with excellent efficiency and good to excellent enantioselectivity Published on 12 June 2014. Downloaded by University of Texas Libraries 18/09/2014 20:37:53. other arylamines, such as 3,4,5-trimethoxyaniline, 3-methoxy- (Scheme 5). Bicyclic products fused with thiophene, pyrrole, aniline, and 3-hydroxyaniline proceeded to give the desired and indole, could also be generated. Oxetanes with two substi- products (22–24) with moderate enantiomeric excess. A tuents at the 3-position could furnish the desired products Figure 1. Nonlinear correlation between the enantiomeric excess of A quaternary chiral center at position 4 of the tetrahydroiso- and that of product 5. with quaternary stereocenters, but unfortunately with moder- ate enantioselectivity. 2028 www.angewandte.org  2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Angew. Chem. Int. Ed. 2013, 52, 2027 –2031 A series of control experiments were carried out to elucidate the reaction mechanism. These experiments suggested that different pathways might be possible depending on substrates (Scheme 6). In path a, hemiaminal 17 was initially formed from the substrates. Rather than imine formation, the nucleo- Scheme 6 Possible mechanisms. philic amine motif opens the oxetane ring, forming cyclic hemiaminal 18. Next, reduction of 18 via iminium 19 affords the observed product 16. Alternatively, in path b, imine 20 is nucleophiles. Encouraged by the successful exploitation of formed followed by reduction to give amine 21. Next, enantio- chiral Brønsted acid catalysis, we next targeted the more chal- selective oxetane ring-opening forms the product 16. It is also lenging intermolecular opening of 3-substituted oxetanes. possible that the imine 20 could undergo intramolecular However, initial screening of different nucleophiles, including oxetane ring-opening to form iminium 19 (path b′). Control commonly used alcohols, amines, and thiols, resulted in no experiments indicated that all these pathways may function. reactions in almost all cases. After considerable effort, we were One of them could be dominant depending on substrates. pleased to find that 2-mercaptobenzothiazoles 23 could The above two reactions represented the first demon- smoothly open the oxetanes 22 in the presence of catalyst B stration of efficient oxetane desymmetrization by nitrogen at room temperature, forming the desired products 24 with

6030 | Org. Biomol. Chem.,2014,12,6028–6032 This journal is © The Royal Society of Chemistry 2014 COMMUNICATION

Table 1. Optimization of reaction conditions.[a] phene (4g), pyrrole (4h-i), and indole (4j) ring can also par- ticipate in the enantioselective cascade process to furnish a diverse set of heterocyclic compounds. Notably, pyrrolopi- perazine and indolopiperazine are both key subunits of a range of important drug leads and bioactive natural prod- ucts, such as longamide B, agelastatin F, and palau Catalytic Asymmetric Ring Opening amine (Scheme 3).[16] In addition, as shown in Table 3, a

§ Chiral Tetrahydroisoquinoline synthesis at C4 position (rare) View Article Online – Hantzsch ester as reductant Perspective Organic & Biomolecular Chemistry – Quant centers lower ee

J. Sun et al.

Table 3. Amine scope.[a] Entry Catalyst Solvent 3 (R1,R2) er (4a) Scheme 3. Representative pyrrolopiperazine-containing alkaloids. 1()-A1 toluene 3a (Et,Et) – Æ 2(R)-A2 toluene 3a 73:27

Entry Ar Product3( Yield [%] erR of)-4 A3 toluene 3a 64:36 4(R)-A4 toluene 3a 88:12 1 3-OH-C6H4 4k 95 96:4 range of aryl amines are suitable reaction partners. The mild 2 3,5-(OMe)2-C6H3 4l 5(96 96:4R)-B toluene 3a 70:30 reaction conditions are compatible with typical functional 3 4-OMe-C6H4 4m 6(89 94:6S)-C1 toluene 3a 92:8 4 4-SMe-C6H4 4n 96 93:7 Scheme 4 Chiral Brønsted acid catalyzed oxetane desymmetrization 7(S)-C2 toluene 3a 84:16 groups, such as ethers, esters, halides, and free alcohols. Fi- 5 4o 95 88:12 with nitrogen nucleophiles. 8[b] (S)-C1 toluene 3a 91:9 nally, the hydroxymethyl group in the 4-position of the prod- 6 2-napthyl 4p 9(96 90:10S)-C1 DCM 3a 70:30 ucts is poised for further functionalizations, such as oxida- [a] Yield of isolated product. phosphoric acid A, we were able to observe the efficient for- 10 (S)-C1 THF 3a 91:9 6 tion and coupling reactions. mation of polycyclic adducts 13 (Scheme 4). After simple puri- 11 (S)-C1 Et2O 3a 93:7 stereogenic center has been already established, compound Scheme 4. Plausible mechanisms. We have carried out several control experiments to probe fication by filtration or centrifugation, these heterocyclic 12 (S)-C1 CPME 3a 94:6 5 was subjected to reduction by Hantzsch ester 3d in the products were isolated in good to excellent enantio- and dia- Scheme 5 Asymmetric syntheses of substituted tetrahydroisoquino- Sun, J. Chem. Eur. J. 2013, 19, 8426.! the reaction mechanism. In the absence of the reductant 3d, presence of racemic phosphoric acid A113. Product 4a was (S ob-)-C1 CPME 3b (tBu,tBu) 90:10 stereopurity. Notably, the reaction generated four new bonds lines and their analogues. tained in quantitative yield and the14 enantioselectivity (S)- isC1 that of theCPME product obtained3c by(t usingBu,Me) the one-pot standard 92:8 the reaction between 1a and 2a affords 5 in quantitative (two C–C and two C–N bonds) and four new stereogenic comparable to that obtained using the one-pot15 standard (S re-)-C1 reaction conditions.CPME 3d (Me,Me) 95:5 centers in one pot from three achiral compounds. Regarding yield as a single diastereomer [Eq. (1)], presumably result- action conditions. However, we could not16 observe[c] any inter-(S)-C1 From theCPME above experimental3d results, we proposed97:3 possi- the reaction mechanism, we proposed the initial formation of mediate, such as 5, under our standard one-pot procedure. ble mechanisms (Scheme 4). As depicted in path a, through ing from formation of the N,O-acetal followed by oxetane the internal amine nucleophile by sequential imine formation These results suggest that if compound [a]5 is The an intermediate, ratio, 1/2/3, ismediation 1:1:1.2; of GC the acid analysis catalyst, using the aldehyde an internal and the standard amine ring opening by attack of the nitrogen atom. Because the C4 and indole addition, followed by oxetane opening. Small posi- the second step might be the faster step.showed In contrast that to all1a products, forms were N,O-acetal obtainedI, an in intermediate quantitative for yield. imine [b] formation. Without tive non-linear effects were also observed, presumably pyrrole-linked substrate 1h could not lead5 Š toMS. a similar [c] Run inter- at Rather208C. than CPME undergoing=cyclopentyl dehydration methyl to form ether, an imine, Tf =I un-tri- suggesting a more complicated mechanism. À mediate. Indeed, no reaction was observedfluoromethanesulfonyl. between 1h and dergoes intramolecular oxetane ring opening through attack In addition to indoles as the reaction partner, we also 2a in the presence of catalyst C1 [Eq. (2)]. However, during of the amine group, which is more nucleophilic than the al-Table 2. Aldehyde scope.[a] employed Hantzsch ester 15 for the above reaction.7 With a the one-pot reaction of 1h under our standard reaction con- cohol group, to give intermediate II. Subsequent reduction similar catalytic system, the reactions of oxetanes 10 and ditions, we were able to observe one intermediate that is ini- of the resulting N,O acetal group, presumably via iminium amines 14 formed a range of tetrahydroisoquinolines 16 with tially generated and then consumed atTable the end. 1, When entry race- 12).III The, affords steric thebulk desired of product, the Hantzsch4. Alternatively, ester the alde- excellent efficiency and good to excellent enantioselectivity mic catalyst A1 was used, the same intermediate, which hyde and the amine may first undergo reductive amination Published on 12 June 2014. Downloaded by University of Texas Libraries 18/09/2014 20:37:53. (Scheme 5). Bicyclic products fused with thiophene, pyrrole, proved to be the reductive aminationalso product has6, was an also effectto on form enantioselectivity. amine V via the imine When intermediate the rela-IV (path b). and indole, could also be generated. Oxetanes with two substi- observed and isolated at partial conversiontively of substratesmall dimethyl as Next, ester, enantioselective3d, is used, ring opening a slight of the increase oxetane delivers an inseparable mixture with racemic product 4h [Eq. (3)]. product 4. It is also possible for imine IV to undergo enan- tuents at the 3-position could furnish the desired products in enantiomeric ratio was observed. Finally, at a lower with quaternary stereocenters, but unfortunately with moder- This mixture was next subjected to the standard reaction tioselective oxetane ring opening to form iminium III conditions, thus forming enantioenriched 4h (81:19 er). The (path b ). According to the experimental results, all these ate enantioselectivity. temperature ( 208C),’ the reaction proceeds with both enantiomeric ratio (er) of 4h that was formed solely in theÀ mechanisms may function, the one dominating being de- A series of control experiments were carried out to elucidate second step was calculated as being 95:5,high which is efficiency similar to pendent and on the excellent substrate. enantioselectivity the reaction mechanism. These experiments suggested that (Table 1, entry 16). To demonstrate the application of our method for THIQ different pathways might be possible depending on substrates With standardsynthesis, reaction we completed conditions a formal established, asymmetric synthesis we of (Scheme 6). In path a, hemiaminal 17 was initially formed the spermidine alkaloid, (+)-(8S,13R)-cyclocelabenzine from the substrates. Rather than imine formation, the nucleo- [17] Scheme 6 Possible mechanisms. next examined the(Scheme scope 5). ofThe the enantioenriched multicomponent product 4m reac-obtained philic amine motif opens the oxetane ring, forming cyclic tion (Table 2). Afrom range our of multicomponent THIQ products reaction was with converted elec- into hemiaminal 18. Next, reduction of 18 via iminium 19 affords phthalimide 7 under Mitsunobu conditions. Subsequent oxi- the observed product 16. Alternatively, in path b, imine 20 is nucleophiles. Encouraged by the successfultron-withdrawing exploitation of dation and electron-donatingat the C1 position delivered groups dihydroisoquinolinone can be 8. formed followed by reduction to give amine 21. Next, enantio- chiral Brønsted acid catalysis, we nextformed targeted the with more chal- highOxidative efficiency cleavage andof the para good-methoxyphenyl to excellent (PMP) group selective oxetane ring-opening forms the product 16. It is also lenging intermolecular opening of 3-substitutedenantioselectivity oxetanes. followed (4a–e by). alkylation A gram-scale afforded 10, reaction a known intermediate was in possible that the imine 20 could undergo intramolecular However, initial screening of different nucleophiles, including a previously completed synthesis of (+)-(8S,13R)-cyclocela- [17b] oxetane ring-opening to form iminium 19 (path b′). Control commonly used alcohols, amines, andalso thiols, carried resulted out, in no andbenzine. no erosion in efficiency and enan- experiments indicated that all these pathways may function. reactions in almost all cases. After considerabletioselectivity effort, we were were observedIn summary, we (4a have), relative developed a to direct the multicomponent small- One of them could be dominant depending on substrates. pleased to find that 2-mercaptobenzothiazoles 23 could method for the efficient assembly of tetrahydroisoquinoline; The above two reactions represented the first demon- smoothly open the oxetanes 22 in theer presence scale of reactions, catalyst B the thus transformation suggesting involves that concomitant our protocol formation is of a C4 stration of efficient oxetane desymmetrization by nitrogen at room temperature, forming the desiredamenable products to24 with large-scalestereocenter multistepby means of enantioselective synthesis. desymmetrization THIQs of oxetanes with amine nucleophiles. Thus, tetrahydroisoqui- with a stereogenicnolines, quaternary privileged center structures at that the are C4 typically position synthesized by a multistep sequence, can now be prepared using a one- 6030 | Org. Biomol. Chem.,2014,12,6028–6032 This journal is © The Royalcan Society also of Chemistry be obtained 2014 in high yield, albeit with moderate step catalytic asymmetric method with high efficiency and enantioselectivity (4f). Substrates having oxetanes teth- [a] Yield of isolated product. [b] Reaction performed at 08C. [c] Reaction performed ered on a heterocyclic aryl ring, for example, on thio- at room temperature. 8428 www.chemeurj.org  2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Chem. Eur. J. 2013, 19, 8426 – 8430

Chem. Eur. J. 2013, 19, 8426 – 8430  2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.chemeurj.org 8427 COMMUNICATION

Table 1. Optimization of reaction conditions.[a] phene (4g), pyrrole (4h-i), and indole (4j) ring can also par- ticipate in the enantioselective cascade process to furnish a diverse set of heterocyclic compounds. Notably, pyrrolopi- perazine and indolopiperazine are both key subunits of a range of important drug leads and bioactive natural prod- ucts, such as longamide B, agelastatin F, and palau Catalytic Asymmetric Ring Opening amine (Scheme 3).[16] In addition, as shown in Table 3, a

§ Chiral TetrahydroisoquinolineSynthesis of Tetrahydroisoquinolines synthesis at C4 position (rare) View Article Online COMMUNICATION – Hantzsch ester as reductant Perspective Organic & Biomolecular Chemistry – Quant centers lower ee Org. Lett. 2012, 14, 3308–3311; h) G. Zhang, Y. Ma, S. Wang, Y. Zhang, R. Wang, J. Am. Chem. Soc. 2012, 134, 12334– 12337. [3] There has been limited success on direct incorporation of a THIQ J. Sunring et al. with a C4 stereogenic center; see, for example: a) L. F. Tietze, K. Thede, R. Schimpf, F. Sannicolò, Chem. Commun. 2000, 583– 584; b) Q.-L. Xu, L.-X. Dai, S.-L. You, Org. Lett. 2012, 14, 2579 – Table 3. Amine scope.[a] 2581; c) D. W. Low, G. Pattison, M. D. Wieczysty, G. H. Churchill, Entry Catalyst Solvent 3 (R1,R2) erH. ( W.4a Lam,) Org. Lett. 2012, 14, 2548– 2551. [4] For some indirectScheme methods, 3. Representative see: a) M. Yamashita, pyrrolopiperazine-containing K. Yamada, K. alkaloids. 1()-A1 toluene 3a (Et,Et) – Æ Tomioka, J. Am. Chem. Soc. 2004, 126, 1954 –1955; b) N. Philippe, 2(R)-A2 toluene 3a 73:27V. Levacher, G. Dupas, G. QuØguiner, J. Bourguignon, Org. Lett. 2000, 2, 2185 –2187; c) M. Kihara, M. Ikeuchi, S. Adachi, Y. Nagao, Entry Ar Product3( Yield [%] erR of)-4 A3 toluene 3a 64:36 4(R)-A4 toluene 3a 88:12H. Moritoki, M. Yamaguchi, Z. Taira, Chem. Pharm. Bull. 1995, 43, 1 3-OH-C6H4 4k 95 96:4 range of aryl amines are suitable reaction partners. The mild 1543– 1546. 2 3,5-(OMe)2-C6H3 4l 5(96 96:4R)-B toluene 3a 70:30 3 4-OMe-C H 4m 89 94:6 [5] A method forreaction forming a C4 conditions stereogenic center are by compatible using an existing with typical functional 6 4 6(S)-C1 toluene 3a 92:8 4 4-SMe-C6H4 4n 96 93:7 isoquinoline skeleton has not been available until recently: L. Shi, Scheme 4 Chiral Brønsted acid catalyzed oxetane desymmetrization 7(S)-C2 toluene 3a 84:16 groups, such as ethers, esters, halides, and free alcohols. Fi- 5 4o 95 88:12 Z.-S. Ye, L.-L. Cao, R.-N. Guo, Y. Hu, Y.-G. Zhou, Angew. Chem. with nitrogen nucleophiles. Scheme 5.8[b] A formal synthesis(S)-C1 of (+)-(8S,13tolueneR)-cyclocelabenzine.3a brsm = 91:9 nally, the hydroxymethyl group in the 4-position of the prod- based on recovered starting material, CAN=cerium ammonium nitrate, 2012, 124, 8411 –8414; Angew. Chem. Int. Ed. 2012, 51, 8286 –8289. 6 2-napthyl 4p 9(96 90:10S)-C1 DCM 3a 70:30 DIAD =diisopropylazodicarboxylate. [6] a) J. W. Cook,ucts J. D. Loudon, is poised In The for Alkaloids, further Vol. functionalizations, 2, (Eds.: R. H. F. such as oxida- ffi [a] Yield of isolated product. phosphoric acid A, we were able to observe the e cient for- 10 (S)-C1 THF 3a 91:9Manske, H. L. Holmes), Academic Press, New York, 1952, p. 331; 6 tion and coupling reactions. mation of polycyclic adducts 13 (Scheme 4). After simple puri- 11 (S)-C1 Et2O 3a 93:7b) R. Apodaca, A. J. Barbier, N. I. Carruthers, L. A. Gomez, J. M. stereogenic center has been already established, compound Scheme 4. Plausible mechanisms. We have carried out several control experiments to probe fication by filtration or centrifugation, these heterocyclic 12 (S)-C1 CPME 3a 94:6Keith, T. W. Lovenberg, R. L. Wolin, PCT Int. Appl. WO 5 was subjected to reduction by Hantzsch ester 3d in the products were isolated in good to excellent enantio- and dia- Scheme 5 Asymmetric synthesesenantioselectivity. of substituted tetrahydroisoquino- The present method can also beSun, applied J. Chem. Eur. J. 2013138604, 19, 8426. A1,! 2006.the reaction mechanism. In the absence of the reductant 3d, presence of racemic phosphoric acid A113. Product 4a was (S ob-)-C1 CPME 3b (tBu,tBu) 90:10 stereopurity. Notably, the reaction generated four new bonds lines and their analogues. to the asymmetric synthesis of other heteroaryl-fused ana- tained in quantitative yield and the14 enantioselectivity (S)- isC1 that of theCPME product obtained3c by(t usingBu,Me) the one-pot[7] standard 92:8 For selected recentthe reaction examples, see: between a) N. T. Tam,1a and C.-G.2a Cho,affordsOrg. 5 in quantitative (two C–C and two C–N bonds) and four new stereogenic comparable to that obtained usinglogs, the such one-pot15 as standard pyrrolopiperazines (S re-)-C1 reaction conditions.CPME and indolopiperazines,3d (Me,Me) 95:5Lett. 2008, 10, 601– 603; b) K. Yamada, M. Yamashita, T. Sumiyoshi, centers in one pot from three achiral compounds. Regarding yield as a single diastereomer [Eq. (1)], presumably result- action conditions. However, wewhich could not are observe[c] also anycore inter- structuresFrom of the many above important experimental alkaloids results, we proposedK. possi- Nishimura, K. Tomioka, Org. Lett. 2009, 11, 1631– 1633; c) C. G. the reaction mechanism, we proposed the initial formation of 16 (S)-C1 CPME 3d 97:3 mediate, such as 5, under ourand standard drug one-pot lead procedure. compounds.ble This mechanisms multicomponent (Scheme 4). As reaction depicted in path a, throughMaÇas, V. L. Paddock,ing from C. G. formation Bochet, A. C. of Spivey, the A. N,O-acetal J. P. White, I. followed by oxetane the internal amine nucleophile by sequential imine formation [a] The ratio, 1/2/3, is 1:1:1.2; GC analysis using an internal standardMann, W. Oppolzer, J. Am. Chem. Soc. 2010, 132, 5176 –5178; These results suggest that if compoundmay proceed5 is an intermediate, by two possiblemediation mechanisms, of the acid either catalyst, of the which aldehyde and the amine ring opening by attack of the nitrogen atom. Because the C4 and indole addition, followed by oxetane opening. Small posi- the second step might be the faster step.showed In contrast that to all1a products, forms were N,O-acetal obtainedI, an in intermediate quantitative for yield. imine [b] formation. Withoutd) A. K. Srivastava, M. Koh, S. B. Park, Chem. Eur. J. 2011, 17, tive non-linear effects were also observed, presumably could be dominant depending on the substrate. Finally, we pyrrole-linked substrate 1h could not lead5 Š toMS. a similar [c] Run inter- at Rather208C. than CPME undergoing=cyclopentyl dehydration methyl to form ether, an imine, Tf4905–=I un-tri- 4913. suggesting a more complicated mechanism. À mediate. Indeed, no reaction washave observed appliedfluoromethanesulfonyl. between our1h protocoland dergoes in the intramolecularformal synthesis oxetane of ring the opening al- through[8] a) attack J. A. Burkhard, G. Wuitschik, M. Rogers-Evans, K. Müller, E. M. In addition to indoles as the reaction partner, we also 2a in the presence of catalyst C1kaloid,[Eq. (2)]. (+ However,)-(8S,13 duringR)-cyclocelabenzine.of the amine group, We which anticipate is more nucleophilic that thanCarreira, the al-TableAngew. 2. Aldehyde Chem. 2010 scope., 122[a], 9236 –9251; Angew. Chem. Int. employed Hantzsch ester 15 for the above reaction.7 With a the one-pot reaction of 1h underthe our present standard method reaction con- will findcohol more group, applications to give intermediate in drugII. dis- Subsequent reductionEd. 2010, 49, 9052 –9067; b) G. Wuitschik, E. M. Carreira, B. similar catalytic system, the reactions of oxetanes 10 and ditions, we were able to observe one intermediate that is ini- of the resulting N,O acetal group, presumably via iminiumWagner, H. Fischer, I. Parrilla, F. Schuler, M. Rogers-Evans, K. covery and alkaloid synthesis. amines 14 formed a range of tetrahydroisoquinolines 16 with tially generated and then consumed atTable the end. 1, When entry race- 12).III The, affords steric thebulk desired of product, the Hantzsch4. Alternatively, ester theM alde-üller, J. Med. Chem. 2010, 53, 3227– 3246. excellent efficiency and good to excellent enantioselectivity mic catalyst A1 was used, the same intermediate, which hyde and the amine may first undergo reductive[9] amination a) W. Zhao, Z. Wang, J. Sun, Angew. Chem. 2012, 124, 6313–6317; Published on 12 June 2014. Downloaded by University of Texas Libraries 18/09/2014 20:37:53. also has an effect on enantioselectivity. When the rela- (Scheme 5). Bicyclic products fused with thiophene, pyrrole, proved to be the reductive amination product 6, was also to form amine V via the imine intermediate IV (pathAngew. b). Chem. Int. Ed. 2012, 51, 6209–6231; b) Z. Chen, B. Wang, observed and isolated at partial conversion of substrate as Next, enantioselective ring opening of the oxetane delivers and indole, could also be generated. Oxetanes with two substi- tively small dimethyl ester, 3d, is used, a slight increaseZ. Wang, G. Zhu, J. Sun, Angew. Chem. 2013, 125, 2081 –2085; an inseparable mixture with racemic product 4h [Eq. (3)]. product 4. It is also possible for imine IV to undergo enan- tuents at the 3-position could furnish the desired products Angew. Chem. Int. Ed. 2013, 52, 2027–2031. This mixture was next subjected to thein standard enantiomeric reaction ratiotioselective was observed.oxetane ringFinally, opening to at form a lower iminium III with quaternary stereocenters, but unfortunately with moder- Acknowledgements [10] For examples of enantioselective desymmetrization of prochiral 3- conditions, thus forming enantioenriched 4h (81:19 er). The (path b’). According to the experimental results, all these ate enantioselectivity. temperature ( 208C), the reaction proceeds with bothsubstituted oxetanes, see: a) M. Mizuno, M. Kanai, A. Iida, K. To- enantiomeric ratio (er) of 4h that was formed solely in theÀ mechanisms may function, the one dominating being de- A series of control experiments were carried out to elucidate Financial support was provided by Hong Kong RGC (GRF604411 and second step was calculated as being 95:5,high which is efficiency similar to pendent and on the excellent substrate. enantioselectivitymioka, Tetrahedron: Asymmetry 1996, 7, 2483 –2484; b) M. Mizuno, the reaction mechanism. These experiments suggested that ECS605812). (Table 1, entry 16). To demonstrate the application of our method forM. THIQ Kanai, A. Iida, K. Tomioka, Tetrahedron 1997, 53, 10699 –10708; different pathways might be possible depending on substrates c) R. N. Loy, E. N. Jacobsen, J. Am. Chem. Soc. 2009, 131, 2786– With standardsynthesis, reaction we completed conditions a formal established, asymmetric synthesis we of (Scheme 6). In path a, hemiaminal 17 was initially formed the spermidine alkaloid, (+)-(8S,13R)-cyclocelabenzine2787; d) Z. Wang, Z. Chen, Angew. Chem. Int. Ed. 2013, DOI: from the substrates. Rather than imine formation, the nucleo- Keywords: alkaloids · asymmetric[17] catalysis · nucleophilic Scheme 6 Possible mechanisms. next examined the(Scheme scope 5). ofThe the enantioenriched multicomponent product 4m reac-obtained10.1002/anie.201300188. philic amine motif opens the oxetane ring, forming cyclic additiontion· organocatalysis (Table 2). A· fromstrained range our of molecules multicomponent THIQ products reaction was with converted elec-[11] S. Searles, into V. P. Gregory, J. Am. Chem. Soc. 1954, 76, 2789– 2790. hemiaminal 18. Next, reduction of 18 via iminium 19 affords phthalimide 7 under Mitsunobu conditions. Subsequent[12] For oxi- leading reviews on chiral Brønsted acid catalysis, see: a) T. the observed product 16. Alternatively, in path b, imine 20 is nucleophiles. Encouraged by the successfultron-withdrawing exploitation of dation and electron-donatingat the C1 position delivered groups dihydroisoquinolinone can beAkiyama,8. J. Itoh, K. Fuchibe, Adv. Synth. Catal. 2006, 348, 999 – formed followed by reduction to give amine 21. Next, enantio- chiral Brønsted acid catalysis, we next targeted the more chal- Oxidative cleavage of the para-methoxyphenyl (PMP)1010; group b) T. Akiyama, Chem. Rev. 2007, 107, 5744–5758; c) M. [1] a) J.formed D. Scott, R. with M. Williams, high efficiencyChem. Rev. 2002 and, 102 good, 1669 –1730; to excellent selective oxetane ring-opening forms the product 16. It is also lenging intermolecular opening of 3-substituted oxetanes. followed by alkylation afforded 10, a known intermediateTerada, in Chem. Commun. 2008, 4097– 4112; d) G. Adair, S. Mukher- b) M.enantioselectivity Chrzanowska, M. D. Rozwadowska, (4a–e). AChem. gram-scale Rev. 2004 reaction, 104, was possible that the imine 20 could undergo intramolecular However, initial screening of different nucleophiles, including a previously completed synthesis of (+)-(8S,13R)-cyclocela-jee, B. List, Aldrichimica Acta 2008, 41, 31– 39; e) D. Kampen, C. M. 3341– 3370. [17b] oxetane ring-opening to form iminium 19 (path b′). Control commonly used alcohols, amines, andalso thiols, carried resulted out, in no andbenzine. no erosion in efficiency and enan- [2] Typical strategies include the Bischler–Napieralski cyclization fol- Reisinger, B. List, Top. Curr. Chem. 2009, 291, 395– 456; f) M. experiments indicated that all these pathways may function. reactions in almost all cases. After considerable effort, we were In summary, we have developed a direct multicomponent lowedtioselectivity by reduction,were the Pictet–Spengler observed (reaction,4a), relative the Pomeranz– to the small-Terada, Synthesis 2010, 1929 –1982; g) M. Rueping, R. M. Koenigs, I. One of them could be dominant depending on substrates. pleased to find that 2-mercaptobenzothiazoles 23 could method for the efficient assembly of tetrahydroisoquinoline;Atodiresei, Chem. Eur. J. 2010, 16, 9350– 9365; h) J. Yu, F. Shi, L. Fritscher cyclization,scale reactions, and the introductionthe thus transformation suggesting of groups involves at that the concomitant C1 our position protocol formation is of a C4 The above two reactions represented the first demon- smoothly open the oxetanes 22 in the presence of catalyst B Gong, Acc. Chem. Res. 2011, 44, 1156– 1171; i) M. Rueping, A. of activated isoquinoline derivativesstereocenter by using by means either of electrophilesenantioselective or desymmetrization stration of efficient oxetane desymmetrization by nitrogen at room temperature, forming the desiredamenable products to24 with large-scale multistep synthesis. THIQsKuenkel, I. Atodiresei, Chem. Soc. Rev. 2011, 40, 4539– 4549. nucleophiles. For a leadingof review oxetanes describing with amine various nucleophiles. methods Thus, for tetrahydroisoqui- [13] a) In ref. [9b], we have demonstrated that a chiral phosphoric acid thewith synthesis a stereogenic of THIQs bearingnolines, quaternary privileged a C1 stereogenic center structures at center, that the are C4 see typically position synthesized can catalyze the formation of a range of complex alkaloid-type mol- ref. [1b]; for more recent examples,by a multistep see: a) sequence, Z. Li, C.-J. can Li, nowOrg. be Lett. prepared using a one- 6030 | Org. Biomol. Chem.,2014,12,6028–6032 This journal is © The Royalcan Society also of Chemistry be obtained 2014 in high yield, albeit with moderate 2004, 6, 4997– 4999; b) M. S.step Taylor, catalytic N. Tokunaga, asymmetric E. method N. Jacobsen, with high efficiencyecules and containing a THIQ unit bearing a C4 stereocenter. However, Angew.enantioselectivity Chem. 2005, 117, 6858 (4f –6862;). SubstratesAngew. Chem. having Int. Ed. oxetanes2005, teth-we believe[a] Yield that of the isolated asymmetric product. induction [b] at Reaction C4 is controlled performed by both at 08C. [c] Reaction performed 44,ered 6700– 6704; on a c) heterocyclicN. Sasamoto, C. Dubs, aryl Y. ring, Hamashima, for example, M. Sodeo- on thio-the catalystat room and temperature. the pre-installed C1 stereocenter. The generation of 8428 www.chemeurj.orgka, J. Am. 2013 Chem. Wiley-VCH Soc. Verlag2006 GmbH, 128 &, Co. 14010 KGaA, –14011; Weinheim d) A. M.Chem. Taylor, Eur. J. 2013, 19, 8426THIQs – 8430 bearing a C4 stereocenter in the absence of a C1 stereocen- S. L. Schreiber, Org. Lett. 2006, 8, 143 –146; e) T. Itoh, M. Miyazaki, ter is highly desirable but much more difficult; b) In ref. [10d], oxe- H. Fukuoka, K. Nagata, A. Ohsawa, Org. Lett. 2006, 8, 1295–1297; tanes are shown to undergo asymmetric ring opening through the f) S.-M.Chem. Lu, Eur. Y.-Q. J. 2013 Wang,, 19, X.-W. 8426 – Han, 8430 Y.-G. Zhou,Angew.2013 Wiley-VCH Chem. Verlagattack GmbH of & thiol Co. nucleophiles. KGaA, Weinheim www.chemeurj.org 8427 2006, 118, 2318– 2321; Angew. Chem. Int. Ed. 2006, 45, 2260 –2263; [14] For reviews on transfer hydrogenation with Hantzsch esters, see: g) F. Berhal, Z. Wu, Z. Zhang, T. Ayad, V. Ratovelomanana-Vidal, a) S. G. Ouellet, A. M. Walji, D. W. C. MacMillan, Acc. Chem. Res.

Chem. Eur. J. 2013, 19, 8426 – 8430  2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.chemeurj.org 8429 View Article Online

Organic & Biomolecular Chemistry Perspective

3. Conclusions

Asymmetric nucleophilic ring-opening of the prochiral 3-sub- stituted oxetanes provides expedient access to a range of highly functionalized three-carbon chiral building blocks. Important progress has been made in the past five years. With (salen)Co(III)-based Lewis acid catalyst systems, the Jacobsen lab realized the first catalytic intramolecular process with internal alcohol nucleophile. More recently, we have success- fully developed a new strategy based on chiral Brønsted acid catalysis and demonstrated highly efficient examples with internal nitrogen nucleophiles as well as the first catalytic intermolecular example using sulphur nucleophiles. However, despite the progress, there remain significant challenges and opportunities in this field. In particular, the intermolecular ring-opening reactions are limited to a very special type of nucleophiles. Future effort should be directed at expanding the nucleophile scope to alcohols and amines as well as weak external nucleophiles, such as halogen- and carbon-based Scheme 7 Intermolecular desymmetrization of 3-substituted oxetanes. nucleophiles. However, the ring-opening alcohol product is a potential competing nucleophile, which should be taken into excellent efficiency as well as enantioselectivity. The mild con- consideration when future new catalytic systems for weak ditions tolerate a wide range of functional groups. Both mono- nucleophiles are designed. Realization of these reactions with and disubstituted oxetanes were suitable substrates and the both high chemical efficiency and good stereocontrol is a products bearing tertiary and quaternary stereocenters were daunting but highly desirable task. formed efficiently. Although the nucleophiles were limited in scope, the arylthio ether moiety in the products could be easily converted to other functional groups (e.g., by Julia olefination). Acknowledgements At present, this reaction is the only example of catalytic asym- metric intermolecular opening of 3-substituted oxetanes We thank HKUST and Hong Kong RGC (GRF-604411Angewandte and (Scheme 7). ECS-605812) for financial support. Chemie Based on mechanistic studies using NMR experiments, we proposed two transition state models for chiral induction. In Table 4: Other suitable nucleophiles. both cases, the primary activation is the hydrogen-bonding Notes and references between the catalyst acidic proton and the basic oxetane Published on 12 June 2014. Downloaded by University of Texas Libraries 18/09/2014 20:37:53. moiety. For oxetanes substituted with a hydrogen-bond donor 1(a) G. Wuitschik, M. Rogers-Evans, K. Müller, H. Fischer, (e.g., OH) at the 3-position, this group provides secondary B. Wagner, F. Schuler, L. Polonchuk and E. M. Carreira, interactionEntry with Ar the phosphoryl oxygenProduct to form Yield the [%][a] nine-mem-ee [%] Angew. Chem., Int. Ed., 2006, 45, 7736–7739; (b) H. Kudo bered transition state TS1 (Fig. 1). However, the orientation of and T. Nishikubo, J. Polym. Sci., Part A: Polym. Chem., 2007, Catalytic Asymmetric Ringthose1 Opening substrates without such a4m hydrogen-bond97 donor substi- 99 45, 709–726; (c) G. Wuitschik, M. Rogers-Evans, A. Buckl, tuent is controlled by steric effect, i.e., the small substituent M. Bernasconi, M. Märki, T. Godel, H. Fischer, B. Wagner, (Rs2) is placed toward the catalyst4n backbone to minimize97 steric 96 FigureI. 1. Parrilla,Plausible transition F. Schuler, states foractivation J. Schneider, of oxetanes A. with Alker, (a) repulsion (Fig. 1b). These models lead the nucleophiles to andW. without B. Schweizer, (b) a hydrogen-bond K. Müller donor and substituent. E. M. Carreira, Angew. View Article Online § Intermolecular Nuc opening of oxetanesapproach with3 common from the back nucleophiles face to open4o the ring, is whichchallenging96 is consist-> 99 ! Chem., Int. Ed., 2008, 47, 4512–4515; (d) J. A. Burkhard, Organic & Biomolecular Chemistry ent with the experimental outcome. PerspectiveG. Wuitschik, M. Rogers-Evans, K. Müller and – Alcohols, amines, and thiols result in mostly no reaction comeE. M. of theCarreira, intermolecularAngew. Chem., ring-opening Int. Ed., process. 2010, 49 We, 9052 believe–9067; 4 4p 73 90 3. Conclusions that(e) the G. hydrogen Wuitschik, bond E. M. between Carreira, the B. catalyst Wagner, OH H. and Fischer, the [a] Yield of purified product. oxetaneI. Parrilla, oxygen F. Schuler, atom is M. the Rogers-Evans primary substrate–catalyst and K. Müller, interaction.J. Med. Chem. Additional, 2010, 53 interactions, 3227–3246. vary with substrates. For Asymmetric[16] nucleophilic ring-opening of the prochiral 3-sub- molecules, but also a versatile functional group which can oxetanes2T.DudevandC.Lim, bearing a hydrogen-bondJ. Am. Chem. Soc. donor,1998, substituent120,4450 (e.g.,–4458. [17] stitutedbe easily oxetanes converted provides into other expedient useful functionalities. access to a rangeOH)3 of For at some the 3-position, examples of we catalytic propose asymmetric a nine-membered ring-openings cyclic highlyThe functionalized enantioenriched three-carbon products obtained chiral by our building catalytic blocks.transitionor ring-expansions state, which involves of 2-substituted another hydrogen oxetanes, bond see: Importantintermolecular progress desymmetrization has been made reactions in the past can five be trans-years. Withbetween(a) H. Nozaki, the catalyst S. Moriuti, phosphoryl H. Takaya oxygen and R. atom Noyori, andTetrahe- the (salen)Co(formed intoIII)-based other useful Lewis compounds acid catalyst (Scheme systems, 2). The the benzo- Jacobsensubstituentdron Lett. (Figure, 1966, 1a).7, 5239 In– contrast,5244; (b) oxetanes H. Nozaki, lacking H. Takaya, such a hydrogen-bond donor substituent may adopt a transition lab realized the first catalytic intramolecular process withS. Moriuti and R. Noyori, Tetrahedron, 1968, 24, 3655–3669; state with the larger substituent (R ) oriented opposite to the Fig.internal 1 Plausible alcohol transition nucleophile. state models. More recently, we have success-(c) K. Ito and T. Katsuki, Chem.L Lett., 1994, 1857–1860; catalyst pocket to minimize steric interactions (Figure 1b). In fully developed a new strategy based on chiral Brønsted acidboth cases, the nucleophile approaches the reactive center catalysis and demonstrated highly efficient examples with This journal is © The Royal Society of Chemistry 2014 from the back side becauseOrg. Biomol. the front Chem.,2014, side is12 blocked,6028–6032 by | the6031 internal nitrogen nucleophiles as well as the first catalyticbulky anthryl group in the catalyst. These rationalizations are intermolecular example using sulphur nucleophiles. However,consistent with the absolute stereochemistry observed for the despite the progress, there remain significant challenges andproduct. opportunities in this field. In particular, the intermolecular Although attempts to obtain a substrate/catalyst cocrystal failed, we were able to observe an obvious substrate–catalyst ring-opening reactions are limited to a very special type of interaction by NMR experiments. As shown in Figure 2, the nucleophiles. Future effort should be directed at expanding the nucleophile scope to alcohols and amines as well as weak external nucleophiles, such as halogen- and carbon-based Scheme 7 Intermolecular desymmetrization of 3-substituted oxetanes. nucleophiles.Scheme 2. Representative However, transformations the ring-opening of the desymmetrization alcohol product is a § Although nucleophiles were limited, they arepotential products.still practical competing nucleophile, which should be taken into ffi excellent e –ciency Convert as well to as other enantioselectivity. products The(i.e. mild Julia con- Olefinationconsideration, etc when) future new catalytic systems for weak ditions tolerate a wide range of functional groups. Both mono- nucleophilesthiazole thioether are designed. can be easily Realization oxidized to of the these corresponding reactions with sulfone, which is widelySun, J. used Angew as. aChem. Julia Int. olefination Ed. 2013 reagent, 52, 6685.! and disubstituted oxetanes were suitable substrates and the both high chemical efficiency and good stereocontrol is a (e.g., formation of 5).[17] Moreover, the primary alcohols can products bearing tertiary and quaternary stereocenters were dauntingalso be oxidized but highly to carbonyl desirable compounds task. with an a-tertiary or ffi formed e ciently. Although the nucleophiles were limited in quaternary chiral center (e.g., aldehyde 6 and ester 7).[18] The scope, the arylthio ether moiety in the products could be easily conventional synthesis of these compounds through the converted to other functional groups (e.g., by Julia olefination). catalytic enantioselective a arylation or a heterofunctionali- Acknowledgements Figure 2. NMR observation of the substrate–catalyst interaction.[14] At present, this reaction is the only example of catalytic asym- zation has been a long-standing topic in organic synthesis and 1 [19] Values on the horizontal axes: d( H) [ppm]. metric intermolecular opening of 3-substituted oxetanes Wesome thank of these HKUST reactions and are Hong still challenging Kong RGC today. (GRF-604411Finally, and enantioenriched 1,1-disubstituted epoxides (e.g., 8) can be (Scheme 7). ECS-605812) for financial support. obtained from the corresponding 1,2-diols after two simple Based on mechanistic studies using NMR experiments, we chemical steps. It is worth noting that 1,1-disubstituted addition of one equivalent of the catalyst C2 to either the proposed two transition state models for chiral induction. In terminal olefins are, in general, challenging substrates for substrate 1a or 1ab results in dramatic shifts of all the proton both cases, the primary activation is the hydrogen-bonding Notesasymmetric and epoxidation. references[20] It is also noteworthy that in all signals in the oxetane ring towards high field. A higher between the catalyst acidic proton and the basic oxetane these transformations no erosion in product enantiopurity catalyst/substrate ratio causes a larger shift, thus indicating Published on 12 June 2014. Downloaded by University of Texas Libraries 18/09/2014 20:37:53. moiety. For oxetanes substituted with a hydrogen-bond donor 1(wasa) observed, G. Wuitschik, thus demonstrating M. Rogers-Evans, that our K. present Müller, method H. Fischer, is a reversible interaction. In addition, more complicated peak (e.g., OH) at the 3-position, this group provides secondary versatileB. Wagner, and complementary F. Schuler, L. to Polonchuk known strategies and E. for M. the Carreira,splitting patterns of the oxetane signals were observed, thus interaction with the phosphoryl oxygen to form the nine-mem- synthesisAngew. of Chem., useful chiralInt. Ed. building, 2006, blocks.45, 7736–7739; (b) H. Kudosuggesting that the interaction imposes a chiral environment As shown in Figure 1, we have proposed possible tran- around the prochiral oxetane ring. These observations fully bered transition state TS1 (Fig. 1). However, the orientation of and T. Nishikubo, J. Polym. Sci., Part A: Polym. Chem., 2007, sition states to rationalize the absolute stereochemical out- agree with our proposed transition states. those substrates without such a hydrogen-bond donor substi- 45, 709–726; (c) G. Wuitschik, M. Rogers-Evans, A. Buckl, ff tuent is controlled by steric e ect, i.e., the small substituent Angew.M. Chem.Bernasconi, Int. Ed. 2013 M., 52, Märki, 6685 –6688 T. Godel, H. Fischer, 2013 Wiley-VCH B. Wagner, Verlag GmbH & Co. KGaA, Weinheim www.angewandte.org 6687 (Rs) is placed toward the catalyst backbone to minimize steric I. Parrilla, F. Schuler, J. Schneider, A. Alker, repulsion (Fig. 1b). These models lead the nucleophiles to W. B. Schweizer, K. Müller and E. M. Carreira, Angew. approach from the back face to open the ring, which is consist- Chem., Int. Ed., 2008, 47, 4512–4515; (d) J. A. Burkhard, ent with the experimental outcome. G. Wuitschik, M. Rogers-Evans, K. Müller and E. M. Carreira, Angew. Chem., Int. Ed., 2010, 49, 9052–9067; (e) G. Wuitschik, E. M. Carreira, B. Wagner, H. Fischer, I. Parrilla, F. Schuler, M. Rogers-Evans and K. Müller, J. Med. Chem., 2010, 53, 3227–3246. 2T.DudevandC.Lim,J. Am. Chem. Soc.,1998,120,4450–4458. 3 For some examples of catalytic asymmetric ring-openings or ring-expansions of 2-substituted oxetanes, see: (a) H. Nozaki, S. Moriuti, H. Takaya and R. Noyori, Tetrahe- dron Lett., 1966, 7, 5239–5244; (b) H. Nozaki, H. Takaya, S. Moriuti and R. Noyori, Tetrahedron, 1968, 24, 3655–3669; Fig. 1 Plausible transition state models. (c) K. Ito and T. Katsuki, Chem. Lett., 1994, 1857–1860;

This journal is © The Royal Society of Chemistry 2014 Org. Biomol. Chem.,2014,12,6028–6032 | 6031 Methodology of Oxetanes

Ring Opening Organometallic Reacons Reacons Termhedron L.etters, Vol. 35, No. 38, pp. 7089-7092. 1994 Pergamon Elsevier Science Ltd Printed in Great Britain oo40-4039/94 $7.oo+o.O0

Aminolysis of Oxetanes: Quite Efficient Catalysis by Lanthanide(III) Trifluoromethansulfonates

Paolo Crotti,* Lucilla Favero, France Macchia, and Mauro Pineschi Lanthanide CatalyzedDipattimeato di Chiiica Bioorganica, UniversitaRing di Piss, via Boaanna Opening 33, 56126 Piss, Italy Dedicated to Prof.G.Berti on the occasion of his 70th birthday

§ Ring opening is similar to oxiranes

§ Strong nucleophilesAbstract: Ln(III)trifluoromethansulfonatesunder basic conditions in CH2Clz efficiently do catalyzenot thenormally aminolysis of trimethylene open oxetanes – Very difficultoxide, to Zoctyl react -, and 2-phenyloxetane,amines atwith r.t., to give oxetanes the corresponding ramino alcohols in very good yields. Although the structure and hybridization of orbitals in oxetanes and oxiranes are largely different, the reactivity of the two systems in the ring opening reactions under acid conditions is similar even if the oxetanes react slightly more slower. 1 Evidently, the lower degree of strain in oxetanes is at least partly § Ln(OTf)3 areoffset great by the greaterpromoters: basicity of the Ybring ,oxygen. Nd ,1 GdOn the contrary, in ring opening reactions carried out in – Reactionthe presence times of ofstrong 2h, nucleophiles r.t. under basic or neutral conditions, oxetanes usually exhibit a scarce reactivity compared with oxiranes. In this sense, it is particularly difficult to obtain the direct reaction of – Yields betweenoxetanes with amines 75-99% even if unhindered (highly oxetanes regioselective and amines are uscd.1~2) By consequence, this reaction cannot be efficiently utilized as a general synthetic method for the preparation of y-aminoalcohols.

HNRPRB R’YoH Ln(OTf), HN, CH2C12, r.t. R2 *3

1, R,=H 4, R,=RpH, R3=C4H9 13, R,=Ph, R2=R3= C,H5 2, RI= C&,7 5, R,=H, R2=CH3, R3=Ph 3, R,=Ph 6, R,=H, RyR3=C,H, 15, R,=Ph, Rs3= 3 7, R,=C,H,,, &=H, R,=PCbH, 8. R,=C,,H,7, R,=H, R,=Ph 17, R,=Ph, Ra3= 9. R,=C8H,7. R2=R3= C2HS 10. R,=CeH,,, R2=R3=X3H9

12, R,=Ph, R2=R3=C2H5

14, R,=Ph, t@=

16, R,=Ph, OS==0

7089 Crotti, C. Tet. Lett. 1994, 35, 7089.! Cobalt Catalyzed C-O Bond Cleavage

§ Hydroformylation of cyclic ethers

O O Co (CO) 2 8 O CO

Eisenmann, J. J. Org. Chem. 1962, 27, 2706.! § Tandem hydroformylation with silanes

O O Co2(CO)8 + HSiEt Me 2 CO MeEt2SiO H 53%

O Co (CO) + HSiEt Me 2 8 H 2 CO MeEt2SiO O 40%

OSiMeEt2 Co2(CO)8 O + HSiEt Me 2 CO H 51% O Dalcanale, E. Synthesis 1986, 492.! Cobalt Catalyzed C-O Bond Cleavage

§ Murai: First catalytic report with oxetanes, previously shown for oxiranes

HSiEt2Me (3 equiv.) O Co2(CO)8 (4% cat.) OSiEt2Me MeEt2SiO + MeEt2SiO CO (1 atm), r.t., 2h R R R R R R

DCM 83 16 Benzene 17 63 n-hexanes trace 96

Proposed Mechanism

HSiR3 + Co2(CO)8

SiR3 O R SiCo(CO ) O Co(CO) 3 4 4 Co(CO) R SiO Co(CO) R SiO 3 -Co(CO) 3 4 3 4 O

HSiR H 3 O SiR3 OSiR3 R3SiO R3SiO Hydride attack

Murai, T. J. Organomet. Chem. 1986, 302, 249.! Cobalt/Ruthenium Carbonylation

§ Oxetane to 5-membered ring carbonylation (harsh conditions) – Carbonylation is regiospecific to least hindered side

§ Cobalt better catalyst for oxetane, ruthenium better for thietane – Thietanes are more reactive – Thietane required both catalysts

Co2(CO)8 (A) (10% cat.) O O Ru3(CO)12 (B) (10% cat.) R O CO (60 atm), DME, 190 °C R

O catalyst yield Me Co2(CO)8 (10% cat.) O O Ru (CO) (10% cat. O A 50 3 12 Me O CO (60 atm), DME, 190 °C, 2d B 20 Ph Me Me Ph A+B 70

O A 40 Insertion takes place with retention of substituent O B 80 stereochemistry A+B 89 n-Hex

O A 0 S B 0 A+B 95 Alper, H. J. Org. Chem. 1989, 54, 21.! Cobalt Carbonylation with N-TMS amines

§ Re / Re / Mn / Fe / Ru / Mo based carbonyl catalysts all failed for this transformation

– Amines induce disproportionation of Co(CO)8 O O Co (CO) (3% cat.) + TMS 2 8 Ph N 120 C Ph N OTMS H ° H 1.5 equiv. CO (1 atm), benzene, 15 h 80%

O O Co (CO) (3% cat.) + TMS 2 8 OTMS Ph N 140 C Ph N H ° H 1.5 equiv. CO (1 atm), benzene, 15 h 66%

O O Co (CO) (3% cat.) + TMS 2 8 Ph N 160 C Ph N OTMS H ° H 1.5 equiv. CO (1 atm), benzene, 15 h 51%

O O Co2(CO)8 (3% cat.) O + TMS O Ph N 160 °C Ph N OTMS H H 30% 1.5 equiv. CO (1 atm), benzene, 15 h

O

Co2(CO)8 (3% cat.) or + No Reaction Ph NH2 120 °C or 160 °C O O CO (1 atm), benzene, 15 h Watanabe, Y. J Chem. Soc., Chem. Commun. 1989, 1253.! C-O Insertion with Iridium

§ Iridium complex known to C-O insert into epoxides – Inserts with β-propiolactone – No O-C-O bond cleavage – Pt(II) and Ni(0) also result in C-O bond cleavage but no metallocycles were obtained

PMe O Formation of Iridalactones 3 Organometallics, Vol. 9, No. 4, 1990 1301 O (C8H18)Ir(PMe3)3Cl Cl O Ir L.^ Table 11. Bond Distances (A) and Angles (deg) for 2 toluene, -30 °C Me3P O Bond Distances Ir(l)-C1(2) 2.4863 (19) P(4)-C(10)PMe 1.82683 (73) Ir(1)-P(3) 2.2432 (22) P(4)-C(ll) 1.8183 (84) Ir(1)-P(4) 2.3364 (19) P(5)-C(12) 1.8231 (76) Ir(l)-P(5) 2.3438 (20) P(5)-C(13) 1.8367 (62) § Mechanism Ir(l)-0(15) 2.1242 (49) P(5)-C(14) 1.8198 (90) D CY / Ir(l)-C(19) 2.1194 (57) 0(15)-C(16) 1.3079 (71) P(3)-C(16) 1.8317 (74) C(16)-0(17) 1.2167 (90) P(3)-C(7) 1.8252 (73) C(16)-C(18) 1.5313 (97) P(3)-C(8) 1.8169 (73) C(l8)-C(l9) 1.5459L (115) O P(4)-C(9) 1.8315 (66) O IrL3Cl Cl O Bond Angles Ir 0(15)-Ir(l)-C(19)Oxidative Addition83.1 (2) C(6)-P(3)-C(7)L 102 (4) c7 O P(5)-Ir(l)-C(l9) 94.3 (2) Ir(l)-P(4)-C(ll) 109.7 (3) c P(5)-1r(1)-0(15) 84.7 (1) Ir(l)-P(4)-C(lO) L 121 (2) P(4)-Ir(l)-C(l9) 93.9 (2) Ir(l)-P(4)-C(9) 116 (3) P(4)-1r(1)-0(15) 86.8 (1) C(lO)-P(4)-C(ll) 103 (3) P(4)-Ir(l)-P(5) 167.5 (1) C(9)-P(4)-C(ll) 101 (3) c3" 1 P(3)-Ir(l)-C(l9) 89.9 (2) C(9)-P(4)-C(lO) 103.1 (3) P(3)-1r(1)-0(15)ClL Ir 172.8 (1)O Ir(l)-P(5)-C(l4) 117 (3) Nucleophilic Cleavage 3 CIO P(3)-1r(l)-P(5) 94.5 (1) Ir(l)-P(5)-C(13) 120 (3) cv P(3)-Ir(l)-P(4) 95O (1) Ir(l)-P(5)-C(l2) 107.6 (2) Cl(2)-Ir(l)-C(l9) 172 (2) C(13)-P(5)-C(14) 104.8 (4) Figure 1. Pluto drawing of a molecule of 2. C1(2)-1r(1)-0(15) 88.4 (1) C(12)-P(5)-C(14) 102 (3) Cl(2)-Ir(l)-P(5) 84.7 (1) C(12)-P(5)-C(13) 104 (3) Reaction of 1 with P-propiolactone can be conveniently Cl(2)-Ir(l)-P(4) 85.8 (1) Ir(l)-0(15)-C(16) 115 (4) followed by 'H NMR (observing lactone disappearance) Cl(2)-Ir(l)-P(3) 98 (1) 0(15)-C(16)-C(18) 115 (6) and by 31P{1H)NMR (observing disappearance of 1 and Ir(l)-P(3)-C(8) 115 (2) 0(15)-C(16)-0(17) 123 (6) Ir(l)-P(3)-C(7) 117 (2) 0(17)-C(16)-C(18) 121 (7) formation of Milstein,2). D. Organometallics 1990, 9, 1300.! Ir(l)-P(3)-C(6) 117 (3) C(16)-C(18)-C(19) 115 (6) Compound 1 undergoes facile dissociation of the cyclo- C(7)-P(3)-C(8) 104 (4) Ir(l)-C(l9)-C(l8) 106.3 (4) octene ligand in solution' (eq 2). Upon addition of the C(6)-P(3)-C(8) 99 (3) ki (CeH14)Ir(PMe3)3CI7 Ir(PMe&CI + CeHI4 (2) several hours for completion) and is much faster than the 2 corresponding reaction of the platinum complex PtMez- (NN) (NN = 1,lO-phenanthroline). Ir(PMe&CI + qo% Ir( qo)(PMed3CI (3) In accordance with this, 1 also undergoes facile oxidative addition of p-butyrolactone, yielding a mixture of products, whereas the above-mentioned platinum complex does not 4 react with this substrate.2a k5 The structure of 2 was determined by 'H, 3'P('H), and Ir(PMe&CI + -2 13C('H)NMR and IR spectroscopies and is confirmed by a single-crystal X-ray diffraction study on crystals grown by slow evaporation of the solvent from a toluene solution. lactone at -30 "C, immediate disappearance of Ir(PMeJ3C1 As expected (Figure 1 and Tables I and 11),a distorted and the lactone takes place to form a new complex (ex- octahedral structure is obtained with the two trans hibiting 31P(1H)NMR 6 -35 (t, J = 16 Hz, 1 P), -52 (d, J phosphines tilted toward the less hindered Ir-0-C plane. = 16 Hz, 2 P). This complex is likely to be a penta- The Ir-P bond trans to PMe3 is significantly longer (by coordinate lactone complex 4 (eq 3). The position of 0.1 A) than that trans to 0, thus reflecting the much larger equilibrium 3 is shifted to the left at higher temperatures trans effect of the phosphine ligand. Interestingly, 2 ex- [K(21 "C) = 21. hibits a considerably shorter C=O and longer C-0 bond Reaction 1 exhibits first-order kinetics in both 1 and the than in the analogous complex PtMe2[CH2CHzC(0)O]- lactone with kobsd(21"C) = 2.3 X L mol-ls-' (in tol- (NN) (NN = 1,lO-phenanthroline)where the two bonds uene-d,). Apparently, the facile equilibria involving cy- are equal in length. This probably reflects a larger con- clooctene dissociation from 1 and lactone coordination do tribution of the resonance form 3 in the case of 2, perhaps not have a significant effect on the overall reaction rate. because of the possibility of back donation to the phos- A typical conversion plot is illustrated in Figure 2. phine ligands. The theoretical lines are derived by using the simple It is noteworthy that exclusive cleavage of the CHz-0 model described by eqs 2-4 in an iterative kinetic modeling rather than the OC-0 bond occurred. Both modes of program, GIT, which combines statistical comparison of nucleophilic cleavage of p-propiolactone are known,6 with experimental data with a GEAR-generated theoretical attack at the alkyl carbon resulting in more strain relief model.* A model in which complex 4 lies in the reaction in the transition state. Reactions of propiolactone with pathway to 2 leads to an unacceptable fit. Pt(Wa and Ni(0)2g also result in alkyl-oxygen bond Other Ir complexes also react with 0-propiolactone, cleavage, although a metallalactone was not isolated in the leading to the following order of reactivity: 1 > Ir(PEk),Cl latter case. > Ir(PMe3)4+PF6->> Ir(PEt3)2(C2H,)zC1> Ir(PMe3)2- (C0)Cl. This indicates (a) higher reactivity for higher electron (6) (a) Searles, S. In Comprehensive Heterocyclic Chemistry; Ka- tritzky, A. R., Rees, C. W., Eds.; Pergamon Press: Oxford, 1984; Vol. 7, density on the metal, as expected for reactions involving pp 381-388. (b) March, J. Advanced Organic Chemistry, 3rd ed.; Wiley: New York, 1985; p 337. (c) Etienne, Y.; Fisher, N. In Heterocyclic Com- pounds with Three- and Four-Membered Rings; Weisserger, A,, Ed.; (7) Milstein, D. To be published. Interscience Publishers: New York, 1964; Vol. 19(2),pp 813-818. (8) Weigert, F. J. Comput. Chem. 1987, 11, 273. Angewandte. Communications

DOI: 10.1002/anie.201203521 Cycloaddition Reactions An Expeditious Route to Eight-Membered Heterocycles By Nickel-

Catalyzed Cycloaddition: Low-Temperature C 2 C 3 Bond Cleavage** sp À sp Puneet Kumar, Kainan Zhang, and Janis Louie*

In the 21st century, chemists have witnessed immense growth prone to polymerization and decomposition.[13] Furthermore, in the field of C H bond activation, which represents an self-condensation of heteroatom-substituted cyclobutanones À elegant method for constructing C C bonds in a manner that occurs under neutral as well as basic reaction conditions. À minimizes waste.[1] Alternatively, C C bond activation pro- Despite these challenges, we successfully developed a Ni/IPr À vides another possible solution. Although significant progress catalyst that can effect the coupling of diynes and azetidi- has been made in the area of C H bond activation, C C bond nones to afford dihydroazocines in excellent yield.[14] This À À activation is still in its infancy.[2] The paucity of developments catalytic method not only provides medium-sized eight- in this area can be attributed to the highly inert nature of C C membered heterocycles that are normally challenging to À s bond and the poor interaction of the orbitals of C C prepare but also represents a remarkable model system for [2a] À s bonds with transition metals. the cleavage of C 2 C 3 bonds at low temperature. sp À sp LowThere Temp is a significant C-C amount Bond of literature Cleavage that describes with Nickel the use of the inherent strain of cyclopropanes (strain 1 energy = 27.6 kcalmolÀ ) in transition-metal-catalyzed reac- §tions. Used[3] Before protected the remarkable amine findingto prevent of Murakami self condensation et al., the 1 §use ofConditions cyclobutanes similar (strain to energyMurakami= 26.4 with kcalmol cyclobutanonesÀ ) in such [4] §reactions Reaction remained run largelyat room unexplored. temp! Since then, appreci- able efforts have been made in using various transition-metal catalysts to harness the latent potential of cyclobutanones.[5] MeO C CO Me Most of these studies2 focused2 on the development of methods O Me O 10 mol% [Ni(cod)2] for accessing carbocycles that were, at the time,20 mol% difficult iPr to At the outset, we reasoned that self-condensation of the synthesize + azetidinone couldCO be2Me minimized through the use of suitable N Our research group has been activeN in developingtoluene, 0 ° nickel-C, 8 h protectingPG groups onCO the2Me nitrogen atom of 3-azetidinones. The Me Me PG catalyzed cycloaddition reactions. Recently, our research reaction betweenMe commercially available 3-Boc-protected group and those of others independently discovered a azetidinone (2a) and malonate diyne 1a was chosen as Entry PG Temp ( °C) Conc. (M) Yield (%) Ni/PPh3-catalyzed method for coupling azetidinones and a model reaction. Given our success in using catalytic 1 Boc r.t. 0.1 70 alkynes to afford 3-piperidones; this method involves cleav- amounts of Ni/PPh3 in toluene as reaction conditions for the age of the C C bond2 attachedBoc to the carbonyl60 of the coupling0.1 of azetidinones35 and alkynes, we initially evaluated [6] À azetidinone. We3 surmised thatBoc if two tethered100 alkynes these0.1 reactions conditions21 for the reaction between 1aand 2a. were employed instead of one, insertion of both of the alkynes In the event, although moderate conversion of azetidinone 2a 4 Boc 0 0.1 84 into the azetidinone C 2 C 3 bond could occur, thus resulting was observed, no desired product was detected (Table 1, sp À sp in the formation of5 eight-memberedBoc N-containing0 heterocy- entry0.05 1). Other88 phosphine ligands were also evaluated clic products [Eq. (1)].6 Medium-sizedTs heterocycles0 are prev- (Table0.05 1, entries 2–8),25 but these reactions also led to little or [7] alent among bioactive7 molecules.Bnh Unfortunately,0 the syn- no0.05 desired product.92 Notably, the reactions where some thesis of eight-membered rings poses a serious challenge desired product was formed were those in which electron- [8–10] because of enthalpic andBnh entropic = factors. In contrast to donating phosphines were used (P(nBu)3, PCy3, and P(Cyp)3 ; cyclobutanone, which was usedPh by thePh research group of Table 1, entries 5—7, respectively). A side reaction that Murakami,[11, 12] heteroatom-substituted cyclobutanones are plaguesLouie, many J. Angew cycloaddition. Chem. Int. Ed. reactions 2012, 51, is 8602. the! oligomerization of alkyne units to give aromatic products.[15] We feared that [*] P. Kumar, K. Zhang, Prof. Dr. J. Louie this reaction was the cause of the low yields and therefore we Department of Chemistry, University of Utah conducted the reaction using Ni/PCy3 as the catalyst and 315 South, 1400 East, Salt Lake City, Utah 84112-0850 (USA) employing slow addition of the diyne (Table 1, entry 6). E-mail: [email protected] Unfortunately, a low yield of product 3aa was still obtained. Homepage: http://www.chem.utah.edu/faculty/louie/index.html We then turned our attention to the highly s-donating [**] We acknowledge the NIH (5R01GM076125) and the NSF (0911017) N-heterocyclic carbene (NHC) ligand, IPr, owing to our for financial support. We thank Dr. J. Muller and Dr. A. Aarif of the previous success in using Ni/IPr in the cycloaddition of both University of Utah for providing HRMS and single-crystal X-ray crystallographic data, respectively. We also thank Ashish Thakur for diyne and enynes and carbonyl compounds such as aldehydes [16] the preparation of diynes 1gand 1h, and for the partial synthesis of and ketones; Murakami and co-workers also used a Ni/IPr 1m. catalyst to facilitate the cycloaddition of diynes and cyclo- [11] Supporting information for this article is available on the WWW butanones. Ultimately, the use of IPr proved to be under http://dx.doi.org/10.1002/anie.201203521. advantageous because the product 3aa was obtained in

8602  2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Angew. Chem. Int. Ed. 2012, 51, 8602 –8606 Angewandte. Communications

A proposed mechanism for the cycloaddition reaction is shown in Scheme 2. Typically, such a cycloaddition would begin with oxidative coupling between an alkyne and the carbonyl group.[11,22,23] In this case, oxidative coupling would

Scheme 1. Products derived from the Ni-catalyzed cycloaddition of diynes 1b–1k and azetidinone 2d. Reaction conditions: azetidinone

2d(1 equiv), diyne 1b–1k (1.2 equiv), [Ni(cod)2] (10 mol%), IPr (20 mol%), toluene, 8h.

products (3jd and 3kd, Scheme 1), which contain a spirocyclic Low Temp C-C Bond Cleavagebackbone, were obtained in good yield. 3-oxacyclobutanone (2e) and diyne 1acould also undergo Angewandte. the Ni/IPr-catalyzed cycloaddition reaction using the standard Communications reaction conditions, thus giving dihydrooxocine 3ae in good yield [Eq. (2)]. Angewandte. Communications

Scheme 2. Proposed mechanism of the cycloaddition reaction.

afford a spirocyclic intermediate M1. Subsequent insertion of

To address the question of regioselectivity when using an the pendant alkyne would give the intermediate M2, which unsymmetrical diyne, diyne 1l, which contains a terminal could then undergo either reductive elimination (to afford alkyne and a methyl-substituted alkyne, was subjected to the products such as 3nd and 3od) or b-carbon elimination to optimized reaction conditions; only one regioisomer was afford metallacycle M . Finally, C C bond forming reductive 3 À observed [3ld; Eq. (3)]. Another unsymmetrical diyne, 1m, elimination occurs to give the dihydroazocine product and the the termini of which differ electronically rather than sterically, catalyst. In the cycloaddition of unsymmetrical diyne 1l, the was also evaluated; again, only one regioisomer was observed regioselectivity is governed by the difference in the size of the [3md, Eq. (3)]. substituents on the alkynes. That is, oxidative coupling AInterestingly, proposed mechanism when diynes for the1n cycloadditionand 1o, which reaction were is between the alkyne bearing the larger group occurs first to shownexpected in Scheme to afford 2. six-membered-ring-fused Typically, such a cycloaddition dihydroazocine would minimize steric interactions between the group and the beginproducts, with were oxidative evaluated, coupling the expected between products an alkyne arising and from the ligand. Indeed, this type of regioselectivity has been found [11,22,23] carbonylC 2 C 3 group.bond cleavageIn were this case, not obtained. oxidative Instead,coupling spirocy- would in various cycloaddition reactions catalyzed by nickel com- sp À sp clic pyranA proposed products mechanism were formed for in the good cycloaddition yields [3nd reactionand 3od, is plexes. However, in the cycloaddition of unsymmetrical diyne [16] Eq.shown (4)]. in SchemeMechanism 2. Typically, is a suchquestion a cycloaddition would 1m, the electronic nature of the alkyne, rather than steric begin with oxidative coupling between an alkyne and the Scheme 1. Products derived from the Ni-catalyzed cycloaddition of8604 www.angewandte.orgcarbonyl group.[11,22,23] In this case, oxidative2012 Wiley-VCH coupling Verlag GmbH would & Co. KGaA, Weinheim Angew. Chem. Int. Ed. 2012, 51, 8602 –8606 Louie, J. Angew. Chem. Int. Ed. 2012, 51, 8602.! diynes 1b–1k and azetidinone 2d. Reaction conditions: azetidinone

2d(1 equiv), diyne 1b–1k (1.2 equiv), [Ni(cod)2] (10 mol%), IPr (20 mol%), toluene, 8h.

Scheme 1. Products derived from the Ni-catalyzed cycloaddition of productsdiynes 1b (3jd–1kandand3kd azetidinone, Scheme2d. 1), Reaction which conditions: contain a spirocyclicazetidinone 2d(1 equiv), diyne 1b–1k (1.2 equiv), [Ni(cod) ] (10 mol%), IPr backbone, were obtained in good yield. 2 (20 mol%), toluene, 8h. 3-oxacyclobutanone (2e) and diyne 1acould also undergo the Ni/IPr-catalyzed cycloaddition reaction using the standard reaction conditions, thus giving dihydrooxocine 3ae in good products (3jd and 3kd, Scheme 1), which contain a spirocyclic yield [Eq. (2)]. backbone, were obtained in good yield. 3-oxacyclobutanone (2e) and diyne 1acould also undergo the Ni/IPr-catalyzed cycloaddition reaction using the standard reaction conditions, thus giving dihydrooxocine 3ae in good yield [Eq. (2)].

Scheme 2. Proposed mechanism of the cycloaddition reaction.

afford a spirocyclic intermediate M1. Subsequent insertion of To address the question of regioselectivity when using an theScheme pendant 2. Proposed alkyne mechanism would give of the the cycloaddition intermediate reaction.M2, which unsymmetrical diyne, diyne 1l, which contains a terminal could then undergo either reductive elimination (to afford alkyne and a methyl-substituted alkyne, was subjected to the products such as 3nd and 3od) or b-carbon elimination to

optimized reaction conditions; only one regioisomer was afford metallacycle M3. Finally, C C bond forming reductive afford a spirocyclic intermediateÀM1. Subsequent insertion of observed [3ld; Eq. (3)]. Another unsymmetrical diyne, 1m, elimination occurs to give the dihydroazocine product and the To address the question of regioselectivity when using an the pendant alkyne would give the intermediate M2, which theunsymmetrical termini of which diyne, differ diyne electronically1l, which rather contains than sterically, a terminal catalyst.could thenIn the undergo cycloaddition either of reductive unsymmetrical elimination diyne (to1l, afford the wasalkyne also evaluated; and a methyl-substituted again, only one alkyne, regioisomer was subjected was observed to the regioselectivityproducts such is as governed3nd and by3od the) difference or b-carbon in the elimination size of the to [3mdoptimized, Eq. (3)]. reaction conditions; only one regioisomer was substituentsafford metallacycle on theM alkynes.. Finally, That C C is, bond oxidative forming coupling reductive 3 À observedInterestingly, [3ld; Eq. when (3)]. diynes Another1n unsymmetricaland 1o, which diyne, were1m, betweenelimination the alkyne occurs bearingto give the the dihydroazocine larger group occurs product first and to the expectedthe termini to afford of which six-membered-ring-fused differ electronically rather dihydroazocine than sterically, minimizecatalyst. steric In the interactions cycloaddition between of unsymmetrical the group diyne and1l the, the products,was also were evaluated; evaluated, again, the only expected one regioisomer products arising was observed from ligand.regioselectivity Indeed, this is governed type of regioselectivity by the difference has in beenthe size found of the C 2 C 3 bond cleavage were not obtained. Instead, spirocy- in various cycloaddition reactions catalyzed by nickel com- sp[3mdÀ sp, Eq. (3)]. substituents on the alkynes. That is, oxidative coupling clic pyranInterestingly, products were when formed diynes in goodand yields [3nd, whichand 3od were, plexes. However, in the cycloaddition of unsymmetrical diyne [16] 1n 1o between the alkyne bearing the larger group occurs first to Eq.expected (4)]. to afford six-membered-ring-fused dihydroazocine 1mminimize, the electronic steric interactions nature of the between alkyne,the rather group than and steric the products, were evaluated, the expected products arising from ligand. Indeed, this type of regioselectivity has been found 8604 www.angewandte.org  2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Angew. Chem. Int. Ed. 2012, 51, 8602 –8606 C 2 C 3 bond cleavage were not obtained. Instead, spirocy- in various cycloaddition reactions catalyzed by nickel com- sp À sp clic pyran products were formed in good yields [3nd and 3od, plexes. However, in the cycloaddition of unsymmetrical diyne Eq. (4)].[16] 1m, the electronic nature of the alkyne, rather than steric

8604 www.angewandte.org  2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Angew. Chem. Int. Ed. 2012, 51, 8602 –8606 Angewandte Chemie

DOI: 10.1002/anie.201306869 Heterocycles Nickel-Catalyzed Cycloaddition of 1,3-DienesC-C bond with 3-Azetidinonesactivation and 3-Oxetanones** Angewandte. Communications Ashish Thakur, Megan E. Facer, and Janis Louie* § Further extension with dienes Table 1: Ligand screening for the nickel-catalyzed cycloaddition of diene [a] Both thermodynamic and kinetic forces make C C bond the synthesis of eight-membered heterocycles by the nickel- 1a and azetidinone 2a. À activation one of the most difficult processes to facilitate.[1] catalyzed insertion of diynes into 3-azetidinones Breaking the strong bond between two carbon atoms is (Scheme 1).[8] further hampered by the increased steric congestion in C C À bonds relative to C H and C X bonds.[1a,b,2] Despite the À À difficulty associated with C C bond activation, a handful of Entry Ligand Conversion [%][b] Yield [%][c] À [1] transition-metal-catalyzed methods have been developed. 1 IPr 83 – In fact, these methods serve as potentially useful technology 2 SIPr 42 – for the construction of complex molecular architectures from 3 IMes 89 – 4 dppf 34 n.d. relatively simple starting materials in a highly atom-econom- 5 dppp – – ical fashion. The C C bond-activation step in these systems À 6 dppb > 99 79 generally occurs by two important processes: 1) oxidative 7 PCy3 25 n.d. addition of the C C bond to the transition metal, or 2) b- 8 PPh3 > 99 75 À 9 P(p-CF C H ) 59 n.d. carbon elimination of transition-metal–alkyl complexes.[1a,j] 3 6 4 3 10 P(p-OMeC6H4)3 70 n.d.

Furthermore, the use of small cyclic systems wherein the 11 P(p-tolyl)3 > 99 79 release of inherent strain provides the necessary driving force [a] Reaction conditions: diene 1a (2 equiv), azetidinone (1 equiv, 0.4m), for C C bond cleavage has been a central theme of C C bond À À [Ni(cod)2] (10 mol%), ligand (20 mol% for entries 1–3; 12 mol% for activation strategies.[1,3,4] entries 4–6; 25 mol% for entries 7–11). [b] The conversion of 1a was Ito and co-workers pioneered the use of cyclobutanones determined by GC with naphthalene as an internal standard. [c] Yield of isolated 3aa. Boc= tert-butoxycarbonyl, cod = 1,5-cyclooctadiene, as substrates in transition-metal-catalyzed reactions involving Cy = cyclohexyl, dppb =1,4-bis(diphenylphosphanyl)butane, dppf= 1,1’- [5] C C bond activation. Since then, a handful of processes Scheme 1. Transition-metal-catalyzed C C bond cleavage in cyclobuta- bis(diphenylphosphanyl)ferrocene, dppp = 1,3-bis(diphenylphospha- Scheme 2. Nickel-catalyzed cycloaddition of 1,3-dienes with azetidi- À À involving the insertion of unsaturated hydrocarbons into the nones, 3-azetidinones, and 3-oxetanones. nyl)propane, IMes =N,N’-bis(2,4,6-trimethylphenyl)imidazol-2-ylidene, none 2a and oxetanone 2b. Reaction conditions: diene 1 (2 equiv), 2a (1 equiv, 0.4m) or 2b (1 equiv, 0.2m), [Ni(cod)2] (10 mol%), P(p-tolyl)3 C C bond of cyclobutanones have been developed. For IPr = N,N’-bis(2,6-diisopropylphenyl)imidazol-2-ylidene, SIPr =N,N’- À bis(2,6-diisopropylphenyl)-4,5-dihydroimidazol-2-ylidene; n.d. =not (25 mol%), 1,4-dioxane, 1008C. [a] A catalyst loading of 15 mol% was example, Murakami and co-workers developed a nickel- determined. required. catalyzed cycloaddition of alkynes and diynes with cyclo- Louie, J. Angew. Chem. Int. Ed. 2013, 52, 12161.! butanones to form six- and eight-membered carbocycles.[6a–d] Despite these recent advances in the C C activation of À have been reported to effectively catalyze the cycloaddition cycloaddition and reacted with azetidinone 2a to afford the Wender et al. reported the rhodium-catalyzed intramolecular cyclobutanones, the intermolecular insertion of 1,3-dienes of diynes and enynes with carbonyl compounds.[8,12] Thus, azocine products 3da and 3ea in good yield. We recently [6+2] cycloaddition of activated cyclobutanones (i.e. 2-vinyl- remains a challenge. Interestingly, Ogoshi et al. have shown azetidinone decomposition in this case could be attributed to reported the nickel-catalyzed cycloaddition of 1,6-diynes and cyclobutanones) and olefins to construct eight-membered that 1,3-dienes do indeed react with cyclobutanones in the the highly basic nature of these ligands in conjunction with the 3-azetidinones to form azocine products with a fused 5,8 ring carbocycles.[6e,f] Murakami and co-workers also demonstrated presence of typical Ni catalysts.[9] The Ni0 complex undergoes high temperature of the reaction. We then turned our focus system.[8] The cycloaddition of the cyclic diene 1f with 3- the use of Rh and Ni catalysts for intramolecular olefin oxidative coupling between the diene and the carbonyl group toward less basic monodentate and bidentate phosphine azetidinone 2a afforded the similar 5,8-ring-fused cyclo- insertion into cyclobutanones to form a variety of carbo- of the cyclobutanone, as is seen in numerous reductive ligands. Whereas dppf and dppp gave poor conversion of adduct 3 fa in good yield and thus complements our diyne– cycles.[6g–k] Furthermore, we and others independently dis- coupling methodologies of unsaturated hydrocarbons and azetidinone (Table 1, entries 4 and 5), the use of dppb led to azetidinone cycloaddition. Our prior attempts to synthesize closed syntheses of substituted 3-piperidones by the nickel- carbonyl compounds (i.e., aldehydes, ketones, carbon dioxide, quantitative conversion and the isolation of product 3aa in heterocycles with fused 6,8 ring systems by the cycloaddition 79% yield (entry 6). Low conversion was observed with the of 1,7-diynes with a 3-azetidinone afforded the spirocyclic catalyzed cycloaddition of alkynes and 3-azacyclobutanones isocyanates, etc.) with a reducing agent.[10, 11] However, in the monodentate ligand PCy (Table 1, entry 7); however, the use pyran products.[8] Gratifyingly, the 6,8-ring-fused azocine 3ga (3-azetidinones).[7] We successfully extended this concept to case of diene substrates, the b-carbon elimination necessary to 3 of PPh3, which has been recently reported to catalyze the was obtained in high yield by the use of diene 1g. Owing to complete a catalytic cycle does not occur, and thus a relatively coupling of alkynes and azetidinones, resulted in isolation of the interest in macrocyclic heterocycles,[14] we synthesized 3 1 [*] A. Thakur, M. E. Facer, Prof. Dr. J. Louie stable catalyst sink results: an h :h -allylalkoxynickel(II) the cycloadduct in 75% yield (Table 1, entry 8). Following the macrocyclic dienes 1h and 1i and subjected them to cyclo- [9] Department of Chemistry, University of Utah complex. Thus, given the difficulties associated with catalyst discovery of PPh3 as a simple and effective ligand, we addition reaction conditions with both azetidinone 2a and 315 South, 1400 East, Salt Lake City, Utah 84112-0850 (USA) turnover, we were delighted to discover effective conditions evaluated other triaryl phosphine ligands for this reaction oxetanone 2b. To our delight, the fused macrocyclic azocines

E-mail: [email protected] for the nickel-catalyzed cycloaddition of azetidinones and 1,3- (Table 1, entries 8–11). The use of P(p-tolyl)3 consistently and oxocines 3ha, 3hb, 3ia, and 3ib were obtained in high Homepage: http://www.chem.utah.edu/directory/faculty/ dienes to afford eight-membered N heterocycles. Herein, we afforded the desired product in high yield.[13] Further yields. The structure of 3hb was determined unambiguously louie.html [15] report these results. optimization led to these final reaction conditions: [Ni(cod)2] by single-crystal X-ray crystallography. Interestingly, the [**] We gratefully acknowledge the NSF (55801742) for financial The nickel-catalyzed cycloaddition was investigated with (10 mol%), P(p-tolyl)3 (25 mol%), 1,4-dioxane, 1008C, 24– reaction between unsymmetrical diene 1j and 3-azetidinone support. We thank Dr. J. Muller and Dr. A. Aarif of the University of the commercially available diene 1a and azetidinone 2a as 48 h. 2a afforded the substituted piperidinone 3ja, rather than the Utah for providing HRMS and single-crystal X-ray crystallographic We explored the scope of this synthetic method under the expected product with an eight-membered azocine ring, in model substrates. Initially, highly s-donating N-heterocyclic data, respectively. We thank E. N. Bess (Sigman group) at the optimized reaction conditions (Scheme 2). The reaction of moderate yield [Eq. (1)]. University of Utah for help with the determination of ee values. We carbenes (NHCs) were explored as potential ligands for this oxetanone 2bwith the volatile diene 1aafforded oxocine 3ab To address the question of regioselectivity in the ring also thank D. J. Tindall for the preparation of dienes 1e and 1j. reaction. However, although good conversion of the azetidi- in moderate yield. Dienes 1b and 1c bearing benzyl and opening of 2-substituted azetidinones, we synthesized 2- Supporting information for this article is available on the WWW none was observed, none of the desired product was detected homobenzyl substituents also underwent cycloaddition with benzyl-3-azetidinone (2c) and treated it under the cyclo- under http://dx.doi.org/10.1002/anie.201306869. (Table 1, entries 1–3). Surprisingly, similar Ni/NHC systems both azetidinone 2a and oxetanone 2b to form the eight- addition conditions with diene 1a[Eq. (2)]. Gratifyingly, only membered N- and O-containing heterocycles 3ba, 3bb, 3ca, the regioisomer 3ac[16] was obtained, in high yield, which and 3cb. Dienes 1d and 1e bearing primary and secondary suggests that preferential cleavage of the C C s bond Angew. Chem. Int. Ed. 2013, 52, 12161 –12165  2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim 12161 À alkyl substituents, respectively, were well-tolerated in this between the carbonyl carbon atom and the unsubstituted

12162 www.angewandte.org  2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Angew. Chem. Int. Ed. 2013, 52, 12161 –12165 Angewandte Chemie

product observed in the nickel-catalyzed cycloaddition of a 3- azetidinone and diphenylacetylene.[7b] The proposed mechanism for this cycloaddition reaction is shown in Scheme 3. The oxidative coupling of the 1,3-diene and the carbonyl group of the 3-azetidinone/3-oxetanone

a carbon atom of the 3-azetidinone takes place to afford the heterocyclic product. Although the reaction was highly regioselective, the product was obtained with only 49% ee, in contrast with our previously reported cycloaddition of alkynes and 3-azetidinones, in which excellent retention of enantiomeric purity was observed when enantiomerically Scheme 3. Proposed mechanism for the nickel-catalyzed cycloaddition pure 2-substituted azetidinones were used.[7a] To rule out the reaction. possibility of nickel-catalyzed racemization of the product, we subjected the chiral eight-membered azocine product (49% ee) to our catalytic conditions [Eq. (3)]. No erosion of would result in the formation of an h3 :h1-allylalkoxynickel(II) [9] complex Z1. This complex would then undergo b-carbon

elimination to afford the intermediate Z2. In the case of 2- substituted azetidinones, b-carbon elimination would occur from the less hindered side of the azetidinone, that is, with cleavage of the CAngewandteC bond between the carbonyl carbon atom À Chemie and the unsubstituted a carbon atom of the 3-azetidinone.

Complex Z2 would then isomerize to complex Z3, which can product observed in the nickel-catalyzedundergo cycloaddition two different of a 3- C(sp3) C(sp3) reductive-elimination [7b] À enantiomeric purity was observedazetidinone over the and course diphenylacetylene. of the pathways to form either the piperidinone or the eight- The proposed mechanism for this cycloaddition reaction is reaction. However, upon the treatment of the enantiomeri- membered heterocyclic product. C C bond-forming reduc- shown in Scheme 3. The oxidative coupling of the 1,3-diene Angewandte3 1 À cally pure 2-benzyl-3-azetidinone 2candwith the the carbonyl Ni catalyst group of in thetive 3-azetidinone/3-oxetanone elimination of the hChemi:h -allylalkylnickel(II)e complex Z3 the absence of the diene but otherwise under the standard generally yields the eight-membered heterocyclic product, as reaction conditions, significant loss of enantiomeric purity observed with dienes 1a–i. However, to explain the formation C-C bond activationwas observed in 24 h [Eq. (4)]. Thisproduct observation observed suggests in the nickel-catalyzed that of the cycloaddition six-membered of a heterocycle 3- 3ja when diene 1jwas used, azetidinone and diphenylacetylene.[7b] in this case reductive elimination must occur at the C3 . The proposed mechanism for this cycloaddition reaction3 1 is Angewandte shown in Scheme 3. The oxidativeposition coupling of of the theh 1,3-diene:h -benzylalkylnickel(II) complex Z3 owing Communications [17] § Ogoshi showed Ni(0) catalysts can do C-C activationand the of carbonyl cyclobutanones group of the andto 3-azetidinone/3-oxetanone stabilization couple with by the phenyl group. dienesTable 1: butLigand resulted screening for in the very nickel-catalyzed stable cycloaddition metallocycles of diene In conclusion, we have developed a Ni/P(p-tol)3-catalyzed 1a and azetidinone 2a.[a] intermolecular cycloaddition of 1,3-dienes and 3-azetidi- § Can form large fuzed heterocycles nones/3-oxetanones. This synthetic method involves C C À activation of the strained four-membered heterocycle to form a carbon atom of the 3-azetidinone takes place to afford the monocyclic and bicyclic eight-membered heterocyclic prod- heterocyclic product. Although the reaction was highly a the chiral 3-azetidinone[b] can undergo[c] C H activation by Ni ucts, which are difficult to access by conventional methods. regioselective,Entry the Ligand product was obtained Conversion with [%] only 49%Yield [%]ee, À with the loss of enantioselectivity [Eq. (5)]. Ho and Aïssa also Interestingly, the use of a diene conjugated with a benzene 1 IPr 83 – in contrast with our previously reported cycloaddition of a 2 SIPrproposed the 42 nickel-catalyzed – C H activation of 3-azetidi- ring led to the formation of a piperidinone rather than an alkynes and 3-azetidinones, in which excellent retention of À 3 IMesnones to explain 89 the formation – of the hydroalkynylation eight-membered heterocycle. Efforts to expand the scope of enantiomeric4 purity dppf was observed 34when enantiomerically n.d. Scheme 3. Proposed mechanism for the nickel-catalyzed cycloaddition pure 2-substituted5 dppp azetidinones were used. –[7a] To rule out – the reaction. the reaction and mechanistic analysis are in progress. a carbon atom of the 3-azetidinone takes place to afford the possibility– 6Regioselectivity of nickel-catalyzed dppb racemization(selective)> 99 of the product, 79 we heterocyclic7 product. PCy3 Although the25 reaction was n.d. highly subjected the chiral eight-membered azocine product 8 PPh3 > 99 75 §regioselective, Drop of ee the is product interesting was obtained with only 49% ee, 3 1 (49% ee)9 to our catalytic P(p-CF3C6H conditions4)3 [Eq.59 (3)]. No erosion n.d. of would result in the formation of an h :h -allylalkoxynickel(II) in contrast10 with ourP(p-OMeC previouslyH ) reported70 cycloaddition n.d. of [9] 6 4 3 complex Z1. This complex wouldExperimental then undergo Sectionb-carbon alkynes11 and 3-azetidinones, P(p-tolyl) in which> 99 excellent retention 79 of 3 elimination to afford the intermediateGeneralZ procedure:. In the case In of a 2-nitrogen-filled glove box, a solution of the enantiomeric purity was observed when enantiomerically 2 [a] Reaction conditions: diene 1a (2 equiv), azetidinone (1 equiv, 0.4m), Scheme 3. Proposed mechanism for thecatalyst nickel-catalyzed (10 mol%; cycloaddition prepared from [Ni(cod)2] and tri(p-tolyl)phos- [7a] substituted azetidinones, b-carbon elimination would occur pure 2-substituted[Ni(cod)2] (10azetidinones mol%), ligand (20 were mol% used. for entriesTo 1–3; rule 12 mol% out the for reaction. from the less hindered side of thephane azetidinone, in a 1:2.5 that molar is, ratio with in 1,4-dioxane) was added to a solution of possibilityentries of nickel-catalyzed 4–6; 25 mol% for entries racemization 7–11). [b] The of conversion the product, of 1a was we determined by GC with naphthalene as an internal standard. [c] Yield of cleavage of the C C bond between the carbonyl carbon atom subjected the chiral eight-membered azocine product À isolated 3aa. Boc= tertAngew.-butoxycarbonyl, Chem. Int. cod Ed.= 1,5-cyclooctadiene,2013, 52, 12161 –12165 and the unsubstituted 2013 Wiley-VCHa carbon Verlag atom GmbH of the & Co. 3-azetidinone. KGaA, Weinheim www.angewandte.org 12163 (49% eeCy) to= cyclohexyl, our catalytic dppb = conditions1,4-bis(diphenylphosphanyl)butane, [Eq. (3)]. No erosion dppf= 1,1 of’- would result in the formation of an h3 :h1-allylalkoxynickel(II) bis(diphenylphosphanyl)ferrocene, dppp = 1,3-bis(diphenylphospha- ComplexScheme 2. Nickel-catalyzedZ2[9]would then cycloaddition isomerize of 1,3-dienes to complex with azetidi-Z3, which can complex Z1. This complex3 would then3 undergo b-carbon nyl)propane, IMes =N,N’-bis(2,4,6-trimethylphenyl)imidazol-2-ylidene, undergonone 2a and two oxetanone different2b. Reaction C(sp conditions:) C(sp ) diene reductive-elimination1 (2 equiv), 2a elimination(1 equiv, 0.4m) to or 2b afford(1 equiv, the 0.2 intermediatem), [Ni(cod)À ] (10Z mol%),. In the P(p-tolyl) case of 2- enantiomericIPr = N,N purity’-bis(2,6-diisopropylphenyl)imidazol-2-ylidene, was observed over the course SIPr = ofN,N the’- pathways to form either the piperidinone2 2 or the3 eight- substituted(25 mol%), 1,4-dioxane, azetidinones, 1008C. [a]b-carbon A catalyst loading elimination of 15 mol% would was occur reaction.bis(2,6-diisopropylphenyl)-4,5-dihydroimidazol-2-ylidene; However, upon the treatment of the enantiomeri- n.d. =not membered heterocyclic product. C C bond-forming reduc- determined. fromrequired. the less hindered side of the azetidinone,À that is, with cally pure 2-benzyl-3-azetidinone 2c with the Ni catalyst in tive elimination of the h3 :h1-allylalkylnickel(II) complex Z cleavage ofLouie, the C JC. Angew bond. betweenChem. Int. the Ed. carbonyl 2013, 52, carbon 12161.! atom3 the absence of the diene but otherwise under the standard generally yields theÀ eight-membered heterocyclic product, as have been reported to effectively catalyze the cycloaddition andcycloaddition the unsubstituted and reacteda withcarbon azetidinone atom of2a theto afford 3-azetidinone. the reaction conditions, significant loss of enantiomeric[8,12] purity observed with dienes 1a–i. However, to explain the formation of diynes and enynes with carbonyl compounds. Thus, Complexazocine productsZ would3da thenand isomerize3ea in good to yield.complex WeZ recently, which can was observed in 24 h [Eq. (4)]. This observation suggests that of the six-membered2 heterocycle 3ja when diene31jwas used, azetidinone decomposition in this case could be attributed to undergoreported the two nickel-catalyzed different C(sp cycloaddition3) C(sp3)reductive-elimination of 1,6-diynes and the highly basic nature of these ligands in conjunction with the in3-azetidinones this case to reductive form azocine elimination productsÀ with must a fused occur 5,8 at ring the C3 enantiomeric purity was observed over the course of the pathways to form either the piperidinone or the eight- high temperature of the reaction. We then turned our focus positionsystem.[8] ofThe the cycloadditionh3 :h1-benzylalkylnickel(II) of the cyclic diene complex1f withZ 3- owing reaction. However, upon the treatment of the enantiomeri- membered heterocyclic product. C C bond-forming3 reduc- toward less basic monodentate and bidentate phosphine toazetidinone stabilization2a afforded by the phenyl the similar group.À 5,8-ring-fused[17] cyclo- cally pure 2-benzyl-3-azetidinone 2c with the Ni catalyst in tive elimination of the h3 :h1-allylalkylnickel(II) complex Z ligands. Whereas dppf and dppp gave poor conversion of adductIn conclusion,3 fa in good we yield have and developed thus complements a Ni/P( ourp-tol) diyne–-catalyzed3 the absenceazetidinone of the (Table diene 1, but entries otherwise 4 and 5), under the use the of dppb standard led to generallyazetidinone yields cycloaddition. the eight-membered Our prior attempts heterocyclic to synthesize product,3 as intermolecular cycloaddition of 1,3-dienes and 3-azetidi- reactionquantitative conditions, conversion significant and loss the of isolation enantiomeric of product purity3aa in observedheterocycles with with dienes fused1a 6,8–i ring. However, systems by to theexplain cycloaddition the formation nones/3-oxetanones. This synthetic method involves C C was observed79% yield in 24 (entry h [Eq. 6). (4)]. Low This conversion observation was observed suggests with that the ofof the 1,7-diynes six-membered with a 3-azetidinone heterocycle 3ja affordedwhen the diene spirocyclic1jwas used,À activation of the[8] strained four-membered heterocycle to form monodentate ligand PCy3 (Table 1, entry 7); however, the use inpyran this products. case reductiveGratifyingly, elimination the 6,8-ring-fused must occur azocine at3ga the C3 of PPh3, which has been recently reported to catalyze the monocyclicwas obtained and in high3 bicyclic1 yield by eight-membered the use of diene 1g heterocyclic. Owing to prod- a position of the h :h -benzylalkylnickel(II)[14] complex Z3 owing the chiralcoupling 3-azetidinone of alkynes can and undergo azetidinones, C resultedH activation in isolation by Ni of ucts,the interest which in are macrocyclic difficult to heterocycles, access by[17] conventionalwe synthesized methods. À to stabilization by the phenyl group. with thethe loss cycloadduct of enantioselectivity in 75% yield [Eq. (Table (5)]. 1, entry Ho and8). Following Aïssa also the Interestingly,macrocyclic dienes the1h useand of1i a dieneand subjected conjugated them with to cyclo- a benzene In conclusion, we have developed a Ni/P(p-tol)3-catalyzed proposeddiscovery the nickel-catalyzed of PPh3 as a C simplea H activation and effective of 3-azetidi- ligand, we ringaddition led reaction to the formation conditions with of a both piperidinone azetidinone rather2a and than an evaluated other triaryl phosphineÀ ligands for this reaction intermolecularoxetanone 2b. To cycloaddition our delight, the fused of 1,3-dienes macrocyclic and azocines 3-azetidi- nones to explain the formation of the hydroalkynylation nones/3-oxetanones.eight-membered heterocycle. This synthetic Efforts method to expand involves the scope C C of (Table 1, entries 8–11). The use of P(p-tolyl)3 consistently and oxocines 3ha, 3hb, 3ia, and 3ib were obtained in high À afforded the desired product in high yield.[13] Further activationtheyields. reaction The of structure the and strained mechanistic of 3hb four-memberedwas analysis determined are unambiguously heterocycle in progress. to form [15] optimization led to these final reaction conditions: [Ni(cod)2] monocyclicby single-crystal and X-ray bicyclic crystallography. eight-memberedInterestingly, heterocyclic the prod- a the chiral(10 3-azetidinone mol%), P(p-tolyl) can3 undergo(25 mol%), C 1,4-dioxane,H activation 100 by8C, Ni 24– ucts,reaction which between are difficult unsymmetrical to access diene by1j conventionaland 3-azetidinone methods. À with the48 loss h. of enantioselectivity [Eq. (5)]. Ho and Aïssa also Interestingly,Experimental2a afforded the the substituted Section use of a piperidinone diene conjugated3ja, rather with than a the benzene We explored the scope of this synthetic method under the expected product with an eight-membered azocine ring, in proposed the nickel-catalyzed Ca H activation of 3-azetidi- ringGeneral led procedure: to the formation In a nitrogen-filled of a piperidinone glove box, rather a solution than of an the optimized reaction conditions (Scheme 2). The reaction of moderate yield [Eq. (1)]. À catalyst (10 mol%; prepared from [Ni(cod) ] and tri(p-tolyl)phos- nones tooxetanone explain2b thewith formation the volatile of diene the1a hydroalkynylationafforded oxocine 3ab eight-memberedTo address the heterocycle. question of Efforts regioselectivity to expand2 in the the ring scope of phane in a 1:2.5 molar ratio in 1,4-dioxane) was added to a solution of in moderate yield. Dienes 1b and 1c bearing benzyl and theopening reaction of 2-substituted and mechanistic azetidinones, analysis we are synthesized in progress. 2- homobenzyl substituents also underwent cycloaddition with benzyl-3-azetidinone (2c) and treated it under the cyclo- Angew. Chem.both Int. azetidinone Ed. 2013, 52, 121612a and –12165 oxetanone 2b to form2013 the Wiley-VCH eight- Verlagaddition GmbH conditions & Co. KGaA, with Weinheim diene 1a[Eq. (2)]. Gratifyingly,www.angewandte.org only 12163 membered N- and O-containing heterocycles 3ba, 3bb, 3ca, the regioisomer 3ac[16] was obtained, in high yield, which and 3cb. Dienes 1d and 1e bearing primary and secondary Experimentalsuggests that preferential Section cleavage of the C C s bond General procedure: In a nitrogen-filled glove box,À a solution of the alkyl substituents, respectively, were well-tolerated in this between the carbonyl carbon atom and the unsubstituted catalyst (10 mol%; prepared from [Ni(cod)2] and tri(p-tolyl)phos- phane in a 1:2.5 molar ratio in 1,4-dioxane) was added to a solution of 12162 www.angewandte.org  2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Angew. Chem. Int. Ed. 2013, 52, 12161 –12165

Angew. Chem. Int. Ed. 2013, 52, 12161 –12165  2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.angewandte.org 12163 Angewandte Chemie

product observed in the nickel-catalyzed cycloaddition of a 3- azetidinone and diphenylacetylene.[7b] The proposed mechanism for this cycloaddition reaction is shown in Scheme 3. The oxidative coupling of theA 1,3-dienengewandte and the carbonyl group of the 3-azetidinone/3-oxetanone Chemie

product observed in the nickel-catalyzed cycloaddition of a 3- azetidinone and diphenylacetylene.[7b] The proposed mechanism for this cycloaddition reaction is shown in Scheme 3. The oxidative coupling of the 1,3-diene and the carbonyl group of the 3-azetidinone/3-oxetanone

a carbon atom of the 3-azetidinone takes place to afford the heterocyclic product. Although the reaction was highly regioselective, the product was obtained with only 49% ee, in contrast with our previously reported cycloaddition of alkynes and 3-azetidinones, in which excellent retention of enantiomeric purity was observed when enantiomerically Scheme 3. Proposed mechanism for the nickel-catalyzed cycloaddition pure 2-substituted azetidinones were used.[7a] To rule out the reaction. possibilitya carbon atom of nickel-catalyzed of the 3-azetidinone racemization takes place of the to product, afford thewe subjectedheterocyclic the product. chiral Although eight-membered the reaction azocine was product highly (49%regioselective,ee) to our the catalytic product conditions was obtained [Eq. (3)]. with No only erosion 49% ee of, would result in the formation of an h3 :h1-allylalkoxynickel(II) [9] in contrast with our previously reported cycloaddition of complex Z1. This complex would then undergo b-carbon

alkynes and 3-azetidinones, in which excellent retention of elimination to afford the intermediate Z2. In the case of 2- enantiomeric purity was observed when enantiomerically substitutedScheme 3. Proposed azetidinones, mechanismb-carbon for the nickel-catalyzed elimination would cycloaddition occur [7a] pure 2-substituted azetidinones were used. To rule out the fromreaction. the less hindered side of the azetidinone, that is, with possibility of nickel-catalyzed racemization of the product, we cleavage of the CAngewandteC bond between the carbonyl carbon atom À Chemie subjected the chiral eight-membered azocine product and the unsubstituted a carbon atom of the 3-azetidinone. (49% ee) to our catalytic conditions [Eq. (3)]. No erosion of would result in the formation of an h3 :h1-allylalkoxynickel(II) Complex Z2 would then isomerize to complex Z3, which can [9] 3 3 product observed in the nickel-catalyzedundergocomplex cycloaddition twoZ1. differentThis of complex a 3- C(sp ) wouldC(sp ) then reductive-elimination undergo b-carbon azetidinone and diphenylacetylene.[7b] À enantiomeric purity was observed over the course of the pathwayselimination to to form afford either the intermediate the piperidinoneZ2. In or the the case eight- of 2- The proposed mechanism for this cycloaddition reaction is reaction. However, upon the treatment of the enantiomeri- memberedsubstituted heterocyclic azetidinones, product.b-carbon C eliminationC bond-forming would reduc- occur shown in Scheme 3. The oxidative coupling of the 1,3-diene from the less hinderedAngewandte side3 1 of theÀ azetidinone, that is, with cally pure 2-benzyl-3-azetidinone 2candwith the the carbonyl Ni catalyst group of in thetive 3-azetidinone/3-oxetanone elimination of the hChemi:h -allylalkylnickel(II)e complex Z3 cleavage of the C C bond between the carbonyl carbon atom the absence of the diene but otherwise under the standard generally yields theÀ eight-membered heterocyclic product, as reaction conditions, significant loss of enantiomeric purity observedand the unsubstituted with dienes 1aa–i.carbon However, atom to explain of the 3-azetidinone. the formation product observed in the nickel-catalyzed cycloaddition of a 3- C-C bond activationwas observed in 24 h [Eq. (4)]. This observation suggests that ofComplex the six-memberedZ2 would then heterocycle isomerize3ja towhen complex dieneZ1j3, whichwas used, can azetidinone and diphenylacetylene.[7b] 3 3 inundergo this case two reductive different eliminationC(sp ) C(sp must) reductive-elimination occur at the C3 The proposed mechanism for this cycloaddition reaction is À Angewandte. enantiomeric purity was observed over the course of the positionpathways of to the formh3 :h1-benzylalkylnickel(II) either the piperidinone complex or theZ owing eight- Communications shown in Scheme 3. The oxidative coupling of the 1,3-diene 3 reaction. However, upon the treatment of the enantiomeri- membered heterocyclic product. C [17]C bond-forming reduc- § Ogoshi showed Ni(0) catalysts can do C-C activationand the of carbonyl cyclobutanones group of the andto 3-azetidinone/3-oxetanone stabilization couple with by the phenyl group.À cally pure 2-benzyl-3-azetidinone 2c with the Ni catalyst in tive elimination of the h3 :h1-allylalkylnickel(II) complex Z dienesTable 1: butLigand resulted screening for in the very nickel-catalyzed stable cycloaddition metallocycles of diene In conclusion, we have developed a Ni/P(p-tol)3-catalyzed3 1a and azetidinone 2athe.[a] absence of the diene but otherwise under the standard intermoleculargenerally yields cycloaddition the eight-membered of 1,3-dienes heterocyclic and product, 3-azetidi- as § Can form large fuzed heterocycles reaction conditions, significant loss of enantiomeric purity nones/3-oxetanones.observed with dienes 1a This–i. However, synthetic to method explain involves the formation C C À wasLoss observed of in ee 24 due h [Eq. to (4)]. Ni This C-H observation Activation! suggests! that activationof the six-membered of the strained heterocycle four-membered3ja when heterocycle diene 1jwas to used,form a carbon atom of the 3-azetidinone takes place to afford the monocyclicin this case and reductive bicyclic elimination eight-membered must heterocyclic occur at the prod- C3 3 1 heterocyclic product. Although the reaction was highly a position of the h :h -benzylalkylnickel(II) complex Z3 owing the chiral 3-azetidinone[b] can undergo[c] C H activation by Ni ucts, which are difficult to access by conventional methods. regioselective,Entry the Ligand product was obtained Conversion with [%] only 49%Yield [%]ee, À [17] with the loss of enantioselectivity [Eq. (5)]. Ho and Aïssa also Interestingly,to stabilization the by use the of phenyl a diene group. conjugated with a benzene 1 IPr 83 – in contrast with our previously reported cycloaddition of a 2 SIPrproposed the 42 nickel-catalyzed – C H activation of 3-azetidi- ringIn led conclusion, to the formation we have developed of a piperidinone a Ni/P(p rather-tol)3-catalyzed than an alkynes and 3-azetidinones, in which excellent retention of À 3 IMesnones to explain 89 the formation – of the hydroalkynylation eight-memberedintermolecular cycloaddition heterocycle. Efforts of 1,3-dienes to expand and the 3-azetidi- scope of enantiomeric4 purity dppf was observed 34when enantiomerically n.d. Scheme 3. Proposed mechanism for the nickel-catalyzed cycloaddition [7a] nones/3-oxetanones. This synthetic method involves C C pure 2-substituted5 dppp azetidinones were used. –To rule out – the reaction. the reaction and mechanistic analysis are in progress. À a carbon6 atom of thedppb 3-azetidinone> takes99 place to afford 79 the activation of the strained four-membered heterocycle to form possibility– Regioselectivity of nickel-catalyzed racemization(selective) of the product, we heterocyclic7 product. PCy3 Although the25 reaction was n.d. highly subjected the chiral eight-membered azocine product monocyclic and bicyclic eight-membered heterocyclic prod- 8 PPh3 > 99 75 §regioselective, Drop of ee the is product interesting was obtained with only 49% ee, a 3 1 (49% ee)9 to our catalytic P(p-CF3Cthe6H conditions4)3 chiral 3-azetidinone [Eq.59 (3)]. No erosion can n.d. undergo of would C resultH activation in the formation by Ni of anucts,h :h which-allylalkoxynickel(II) are difficult to access by conventional methods. in contrast with our previously reported cycloaddition of À [9] 10 P(p-OMeC6H4)3 70 n.d. Experimental Section with the loss of enantioselectivity [Eq.complex (5)]. HoZ1. andThis Aï complexssa also wouldInterestingly, then undergo theb-carbon use of a diene conjugated with a benzene alkynes11 and 3-azetidinones, P(p-tolyl) in which> 99 excellent retention 79 of 3 a elimination to afford the intermediateGeneralZ procedure:. In the case In of a 2-nitrogen-filled glove box, a solution of the enantiomeric purity wasproposed observed the when nickel-catalyzed enantiomerically C H activation of 3-azetidi- ring led2 to the formation of a piperidinone rather than an [a] Reaction conditions: diene 1a (2 equiv), azetidinone (1 equiv, 0.4m), ÀScheme 3. Proposed mechanism for thecatalyst nickel-catalyzed (10 mol%; cycloaddition prepared from [Ni(cod)2] and tri(p-tolyl)phos- [7a] substituted azetidinones, b-carbon elimination would occur pure 2-substituted[Ni(cod)2] (10azetidinones mol%),nones ligand (20 were to mol% explain used. for entriesTo the 1–3; rule 12 formation mol% out the for reaction. of the hydroalkynylation eight-membered heterocycle. Efforts to expand the scope of from the less hindered side of thephane azetidinone, in a 1:2.5 that molar is, ratio with in 1,4-dioxane) was added to a solution of possibilityentries of nickel-catalyzed 4–6; 25 mol% for entries racemization 7–11). [b] The of conversion the product, of 1a was we the reaction and mechanistic analysis are in progress. determined by GC with naphthalene as an internal standard. [c] Yield of cleavage of the C C bond between the carbonyl carbon atom subjected the chiral eight-membered azocine product À isolated 3aa. Boc= tertAngew.-butoxycarbonyl, Chem. Int. cod Ed.= 1,5-cyclooctadiene,2013, 52, 12161 –12165 and the unsubstituted 2013 Wiley-VCHa carbon Verlag atom GmbH of the & Co. 3-azetidinone. KGaA, Weinheim www.angewandte.org 12163 (49% eeCy) to= cyclohexyl, our catalytic dppb = conditions1,4-bis(diphenylphosphanyl)butane, [Eq. (3)]. No erosion dppf= 1,1 of’- would result in the formation of an h3 :h1-allylalkoxynickel(II) bis(diphenylphosphanyl)ferrocene, dppp = 1,3-bis(diphenylphospha- ComplexScheme 2. Nickel-catalyzedZ2[9]would then cycloaddition isomerize of 1,3-dienes to complex with azetidi-Z3, which can complex Z1. This complex3 would then3 undergo b-carbon nyl)propane, IMes =N,N’-bis(2,4,6-trimethylphenyl)imidazol-2-ylidene, undergonone 2a and two oxetanone different2b. Reaction C(sp conditions:) C(sp ) diene reductive-elimination1 (2 equiv), 2a elimination(1 equiv, 0.4m) to or 2b afford(1 equiv, the 0.2 intermediatem), [Ni(cod)À Experimental] (10Z mol%),. In the P(p-tolyl) case Section of 2- enantiomericIPr = N,N purity’-bis(2,6-diisopropylphenyl)imidazol-2-ylidene, was observed over the course SIPr = ofN,N the’- pathways to form either the piperidinone2 2 or the3 eight- bis(2,6-diisopropylphenyl)-4,5-dihydroimidazol-2-ylidene; n.d. =not substituted(25 mol%), 1,4-dioxane, azetidinones, 1008C. [a]b-carbon A catalystGeneral loading elimination of procedure: 15 mol% would was In occur a nitrogen-filled glove box, a solution of the reaction. However, upon the treatment of the enantiomeri- memberedrequired. heterocyclic product. C C bond-forming reduc- determined. from the less hindered side of thecatalyst azetidinone,À (10 mol%; that is, prepared with from [Ni(cod)2] and tri(p-tolyl)phos- cally pure 2-benzyl-3-azetidinone 2c with the Ni catalyst in tive elimination of the h3 :h1-allylalkylnickel(II) complex Z cleavage ofLouie, the C JC. Angew bond. betweenChem. Int.phane the Ed. carbonyl 2013 in a, 1:2.5 52, carbon 12161. molar! atom ratio3 in 1,4-dioxane) was added to a solution of the absence of the diene but otherwise under the standard generally yields theÀ eight-membered heterocyclic product, as have been reported to effectively catalyze the cycloaddition andcycloaddition the unsubstituted and reacteda withcarbon azetidinone atom of2a theto afford 3-azetidinone. the reaction conditions, significant loss of enantiomeric[8,12] purity observed with dienes 1a–i. However, to explain the formation of diynes and enynes with carbonyl compounds. Thus, Complexazocine productsZ would3da thenand isomerize3ea in good to yield.complex WeZ recently, which can was observed in 24 h [Eq.Angew. (4)]. Chem. This observation Int. Ed. 2013, suggests52, 12161 –12165 that of the six-membered2  2013 Wiley-VCHheterocycle Verlag3ja GmbHwhen diene & Co.3 KGaA,1jwas Weinheim used, www.angewandte.org 12163 azetidinone decomposition in this case could be attributed to undergoreported the two nickel-catalyzed different C(sp cycloaddition3) C(sp3)reductive-elimination of 1,6-diynes and the highly basic nature of these ligands in conjunction with the in3-azetidinones this case to reductive form azocine elimination productsÀ with must a fused occur 5,8 at ring the C3 enantiomeric purity was observed over the course of the pathways to form either the piperidinone or the eight- high temperature of the reaction. We then turned our focus positionsystem.[8] ofThe the cycloadditionh3 :h1-benzylalkylnickel(II) of the cyclic diene complex1f withZ 3- owing reaction. However, upon the treatment of the enantiomeri- membered heterocyclic product. C C bond-forming3 reduc- toward less basic monodentate and bidentate phosphine toazetidinone stabilization2a afforded by the phenyl the similar group.À 5,8-ring-fused[17] cyclo- cally pure 2-benzyl-3-azetidinone 2c with the Ni catalyst in tive elimination of the h3 :h1-allylalkylnickel(II) complex Z ligands. Whereas dppf and dppp gave poor conversion of adductIn conclusion,3 fa in good we yield have and developed thus complements a Ni/P( ourp-tol) diyne–-catalyzed3 the absenceazetidinone of the (Table diene 1, but entries otherwise 4 and 5), under the use the of dppb standard led to generallyazetidinone yields cycloaddition. the eight-membered Our prior attempts heterocyclic to synthesize product,3 as intermolecular cycloaddition of 1,3-dienes and 3-azetidi- reactionquantitative conditions, conversion significant and loss the of isolation enantiomeric of product purity3aa in observedheterocycles with with dienes fused1a 6,8–i ring. However, systems by to theexplain cycloaddition the formation nones/3-oxetanones. This synthetic method involves C C was observed79% yield in 24 (entry h [Eq. 6). (4)]. Low This conversion observation was observed suggests with that the ofof the 1,7-diynes six-membered with a 3-azetidinone heterocycle 3ja affordedwhen the diene spirocyclic1jwas used,À activation of the[8] strained four-membered heterocycle to form monodentate ligand PCy3 (Table 1, entry 7); however, the use inpyran this products. case reductiveGratifyingly, elimination the 6,8-ring-fused must occur azocine at3ga the C3 of PPh3, which has been recently reported to catalyze the monocyclicwas obtained and in high3 bicyclic1 yield by eight-membered the use of diene 1g heterocyclic. Owing to prod- a position of the h :h -benzylalkylnickel(II)[14] complex Z3 owing the chiralcoupling 3-azetidinone of alkynes can and undergo azetidinones, C resultedH activation in isolation by Ni of ucts,the interest which in are macrocyclic difficult to heterocycles, access by[17] conventionalwe synthesized methods. À to stabilization by the phenyl group. with thethe loss cycloadduct of enantioselectivity in 75% yield [Eq. (Table (5)]. 1, entry Ho and8). Following Aïssa also the Interestingly,macrocyclic dienes the1h useand of1i a dieneand subjected conjugated them with to cyclo- a benzene In conclusion, we have developed a Ni/P(p-tol)3-catalyzed proposeddiscovery the nickel-catalyzed of PPh3 as a C simplea H activation and effective of 3-azetidi- ligand, we ringaddition led reaction to the formation conditions with of a both piperidinone azetidinone rather2a and than an evaluated other triaryl phosphineÀ ligands for this reaction intermolecularoxetanone 2b. To cycloaddition our delight, the fused of 1,3-dienes macrocyclic and azocines 3-azetidi- nones to explain the formation of the hydroalkynylation nones/3-oxetanones.eight-membered heterocycle. This synthetic Efforts method to expand involves the scope C C of (Table 1, entries 8–11). The use of P(p-tolyl)3 consistently and oxocines 3ha, 3hb, 3ia, and 3ib were obtained in high À afforded the desired product in high yield.[13] Further activationtheyields. reaction The of structure the and strained mechanistic of 3hb four-memberedwas analysis determined are unambiguously heterocycle in progress. to form [15] optimization led to these final reaction conditions: [Ni(cod)2] monocyclicby single-crystal and X-ray bicyclic crystallography. eight-memberedInterestingly, heterocyclic the prod- a the chiral(10 3-azetidinone mol%), P(p-tolyl) can3 undergo(25 mol%), C 1,4-dioxane,H activation 100 by8C, Ni 24– ucts,reaction which between are difficult unsymmetrical to access diene by1j conventionaland 3-azetidinone methods. À with the48 loss h. of enantioselectivity [Eq. (5)]. Ho and Aïssa also Interestingly,Experimental2a afforded the the substituted Section use of a piperidinone diene conjugated3ja, rather with than a the benzene We explored the scope of this synthetic method under the expected product with an eight-membered azocine ring, in proposed the nickel-catalyzed Ca H activation of 3-azetidi- ringGeneral led procedure: to the formation In a nitrogen-filled of a piperidinone glove box, rather a solution than of an the optimized reaction conditions (Scheme 2). The reaction of moderate yield [Eq. (1)]. À catalyst (10 mol%; prepared from [Ni(cod) ] and tri(p-tolyl)phos- nones tooxetanone explain2b thewith formation the volatile of diene the1a hydroalkynylationafforded oxocine 3ab eight-memberedTo address the heterocycle. question of Efforts regioselectivity to expand2 in the the ring scope of phane in a 1:2.5 molar ratio in 1,4-dioxane) was added to a solution of in moderate yield. Dienes 1b and 1c bearing benzyl and theopening reaction of 2-substituted and mechanistic azetidinones, analysis we are synthesized in progress. 2- homobenzyl substituents also underwent cycloaddition with benzyl-3-azetidinone (2c) and treated it under the cyclo- Angew. Chem.both Int. azetidinone Ed. 2013, 52, 121612a and –12165 oxetanone 2b to form2013 the Wiley-VCH eight- Verlagaddition GmbH conditions & Co. KGaA, with Weinheim diene 1a[Eq. (2)]. Gratifyingly,www.angewandte.org only 12163 membered N- and O-containing heterocycles 3ba, 3bb, 3ca, the regioisomer 3ac[16] was obtained, in high yield, which and 3cb. Dienes 1d and 1e bearing primary and secondary Experimentalsuggests that preferential Section cleavage of the C C s bond General procedure: In a nitrogen-filled glove box,À a solution of the alkyl substituents, respectively, were well-tolerated in this between the carbonyl carbon atom and the unsubstituted catalyst (10 mol%; prepared from [Ni(cod)2] and tri(p-tolyl)phos- phane in a 1:2.5 molar ratio in 1,4-dioxane) was added to a solution of 12162 www.angewandte.org  2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Angew. Chem. Int. Ed. 2013, 52, 12161 –12165

Angew. Chem. Int. Ed. 2013, 52, 12161 –12165  2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.angewandte.org 12163 Minireviews M. Rogers-Evans et al.

Copper Catalyzed4.3. Ring-Expansion Ring Reactions Expansions of Oxetanes to THF

As part of their studies of transition-metal-catalyzed § First discoveredenantioselective by Noyori in 1966 reactions, Noyori and co-workers document- ed the copper-catalyzedPh asymmetric ring expansion of ox- Me [66] etanes to afford tetrahydrofurans.O They found that the Scheme 16. Selective opening of 2-phenyloxetane with aryl borates. N insertion of a salicylaldimine-chelatedCu copper carbenoid into O N the CO O bond can effect the formation of enantiomerically Ph À Me CO2Et CO2Et O enriched substituted tetrahydrofurans.Ph Inspired by this work, + O O + O 4.2. Intramolecular Ring-Opening Reactions Katsuki and co-workersneat, were 60 ° interestedC in furtherPh improving Ph N2 the reaction. The use of C2-symmetric bipyridine ligands for Bach et al. examined the intramolecular opening of 2,3,3- copper(I) was advantageous for the generation85% of enantioen-yield low ee trisubstituted oxetanes to give ring-expanded ethers, thioeth- riched products (Scheme 18).[67] Treatment of racemic 2- Noyori, R. Tet. Lett. 1966, 7, 5239.! ers, and carbonates.[63] Loy and Jacobsen documented the phenyloxetane (18) with tert-butyl diazoacetate and a mixture enantioselective Lewis acid mediated intramolecular§ Katsuki ringswitchedof 1to mol% bipyridine of CuOTf ligands and bipyridineto enhance ligand ee 100 afforded the opening of oxetanes.[64] Achiral 3,3-disubstituted oxetanes bearing a remote hydroxy group were treated at ambient temperature either neat or at high concentrations (for example, 6m in CH3CN) with catalytic amounts of cobalt(III) salen complexes. The ring-opening reactions afforded a variety of tetrahydrofurans 95 in high yield and enantiose- lectivity (Scheme 17). In accordance with epoxide-opening reactions,[65] dimeric complexes 97 (n = 1) were found to be

Scheme 18. Asymmetric ring expansion of oxetanes according to Katsuki, T. Chem. Lett. 1994, 1857.! Katsuki and co-workers.[67] TBS =tert-butyldimethylsilyl.

products 98 and 99 in a diastereomeric ratio of 59:41 in favor of the trans isomer.[67a,b] The conversion of 2-alkynyloxetane into the corresponding tetrahydrofuran proved to be superior, as the product was formed in 88% yield (as a cis/trans mixture). This transformation was applied in the total syn- thesis of the trans-Whisky lactone and the formal total syntheses of ( )-avenaciolide and ( )-isoavenaciolide.[67c–e] À À In 2001, Lo and Fu described the use of a copper(I) bis(azaferrocene) complex in an oxetane ring-expansion reaction.[68] By using enantiomerically highly enriched 2- substituted oxetanes and a diazoacetate with 1 mol% of a copper(I) catalyst, the ring-expanded products could be obtained in good yield as well as high diastereomeric and enantiomeric ratios (Scheme 19). Sterically demanding car- Scheme 17. Cobalt(III)-catalyzed enantioselective intramolecular open- ing of oxetanes according to Loy and Jacobsen.[64] OTf =trifluoro- methane sulfonate. superior to the monomeric cobalt(III) salen counterpart (96). Thus, the use of catalysts of type 97 allowed catalyst loadings as little as 0.01 mol% to be used, with the rearranged products formed in up to 98% yield and 99% ee. The enhanced reactivity of catalyst 97 (n = 1) can be explained by cooperative bimetallic mechanisms involving simultaneous activation of the nucleophile as well as Lewis acid activation of the oxetane.

Scheme 19. Ring expansion of oxetanes according to Lo and Fu.[68] Cy = cyclohexyl.

9060 www.angewandte.org  2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Angew. Chem. Int. Ed. 2010, 49, 9052 – 9067 Minireviews M. Rogers-Evans et al.

4.3. Ring-Expansion Reactions of Oxetanes

As part of their studies of transition-metal-catalyzed Minireviews M. Rogers-Evansenantioselective et al. reactions, Noyori and co-workers document- ed the copper-catalyzed asymmetric ring expansion of ox- etanes to afford tetrahydrofurans.[66] They found that the Scheme 16. Selective4.3. Ring-Expansion opening of 2-phenyloxetane Reactions with aryl of Oxetanes borates. insertion of a salicylaldimine-chelated copper carbenoid into the C O bond can effect the formation of enantiomerically As part of their studies of transition-metal-catalyzedÀ enriched substituted tetrahydrofurans. Inspired by this work, enantioselective reactions, Noyori and co-workers document- 4.2. Intramolecular Ring-Opening Reactions Katsuki and co-workers were interested in further improving ed the copper-catalyzed asymmetric ring expansion of ox- the reaction. The use of C -symmetric bipyridine ligands for etanes to afford tetrahydrofurans.[66] They found that the 2 Scheme 16. Selective opening of 2-phenyloxetane with aryl borates.Bach et al. examined the intramolecular opening of 2,3,3- copper(I) was advantageous for the generation of enantioen- insertion of a salicylaldimine-chelated copper carbenoid into trisubstituted oxetanes to give ring-expanded ethers, thioeth- riched products (Scheme 18).[67] Treatment of racemic 2- the C O bond can effect the formation of enantiomerically ers, and carbonates.À [63] Loy and Jacobsen documented the phenyloxetane (18) with tert-butyl diazoacetate and a mixture enriched substituted tetrahydrofurans. Inspired by this work, enantioselective Lewis acid mediated intramolecular ring of 1 mol% of CuOTf and bipyridine ligand 100 afforded the 4.2. Intramolecular Ring-Opening Reactions Katsuki and co-workers were interested in further improving opening of oxetanes.[64] Achiral 3,3-disubstituted oxetanes the reaction. The use of C -symmetric bipyridine ligands for bearing a remote hydroxy group were2 treated at ambient Bach et al. examined the intramolecular opening of 2,3,3- copper(I) was advantageous for the generation of enantioen- temperature either neat or at high concentrations (for trisubstituted oxetanes to give ring-expanded ethers, thioeth- riched products (Scheme 18).[67] Treatment of racemic 2- example, 6m in CH CN) with catalytic amounts of cobalt(III) ers, and carbonates.[63] Loy and Jacobsen documented the phenyloxetane3 (18) with tert-butyl diazoacetate and a mixture salen complexes. The ring-opening reactions afforded a enantioselective Lewis acid mediated intramolecular ring of 1 mol% of CuOTf and bipyridine ligand 100 afforded the variety of tetrahydrofurans 95 in high yield and enantiose- opening of oxetanes.[64] Achiral 3,3-disubstituted oxetanes lectivity (Scheme 17). In accordance with epoxide-opening bearing a remote hydroxy group were treated at ambient reactions,[65] dimeric complexes 97 (n = 1) were found to be temperature either neat or at high concentrations (for example, 6m in CH3CN) with catalytic amounts of cobalt(III) salen complexes. The ring-opening reactions afforded a variety of tetrahydrofurans 95 in high yield and enantiose- Scheme 18. Asymmetric ring expansion of oxetanes according to lectivity (Scheme 17). In accordance with epoxide-opening Katsuki and co-workers.[67] TBS =tert-butyldimethylsilyl. reactions,[65] dimeric complexes 97 (n = 1) were found to be products 98 and 99 in a diastereomeric ratio of 59:41 in favor of the trans isomer.[67a,b] The conversion of 2-alkynyloxetane Scheme 18. Asymmetric ring expansion of oxetanes accordinginto to the corresponding tetrahydrofuran proved to be superior, CopperKatsuki and Catalyzed co-workers.[67] TBS =tert-butyldimethylsilyl. Ring Expansionsas the product was to formed THF in 88% yield (as a cis/trans mixture). This transformation was applied in the total syn- thesis of the trans-Whisky lactone and the formal total products 98 and 99 in a diastereomeric ratio of 59:41 in favor syntheses of ( )-avenaciolide and ( )-isoavenaciolide.[67c–e] § Transformationof the trans isomer. applied[67a,b] Thetotal conversion synthesis of of 2-alkynyloxetane trans-Whisky lactoneÀ and formal total synthesisÀ In 2001, Lo and Fu described the use of a copper(I) ofinto (−)- theavenaciolide corresponding and tetrahydrofuran (−)-isoavenaciolide proved to be superior, bis(azaferrocene) complex in an oxetane ring-expansion as the product was formed in 88% yield (as a cis/trans reaction.[68] By using enantiomerically highly enriched 2- mixture). This transformation wasN appliedN in the total syn- substitutedn-C oxetanes7H15 and an-C diazoacetate7H15 with 1 mol% of a thesis of then-C Htrans-Whisky lactoneCu and the formal total 7 15 OTf (1 mol%)copper(I) catalyst, the ring-expanded products could be syntheses of ( )-avenaciolideO and ( )-isoavenaciolide.[67c–e] À OTBSÀ TBSO obtained in good yield as well as high diastereomeric and In 2001,+ Lo andOtBu Fu described the use of a copper(I) + O chlorobenzene, 0 C enantiomericCO ratiostBu (SchemeCO 19).tBu Sterically demanding car- bis(azaferrocene)N complex in an oxetane ring-expansion° 2 2 2 69% yield O O Scheme 17. Cobalt(III)-catalyzedreaction.93% ee [68] By enantioselective using enantiomerically intramolecular highly open- enriched 2- [64] ing of oxetanessubstituted according to oxetanes Loy and Jacobsen. and a diazoacetateOTf =trifluoro- with 1 mol% of85 a : 15 methane sulfonate. 72% ee copper(I) catalyst, the ring-expanded products could be Katsuki, T. Syn. Lett. 1997, 387.! § Bestobtained expansion in good to yielddate asis wellfrom asFu highʼs lab diastereomeric in 2001 and enantiomeric ratios (Scheme 19). Sterically demanding car- superior to the monomeric cobalt(III) salen counterpart (96). Scheme 17. Cobalt(III)-catalyzed enantioselective intramolecularThus, open- the use of catalysts of type 97 allowed catalyst loadings [64] ing of oxetanes according to Loy and Jacobsen. OTf =trifluoro-as little as 0.01 mol% to be used, with the rearranged methane sulfonate. products formed in up to 98% yield and 99% ee. The enhanced reactivity of catalyst 97 (n = 1) can be explained by cooperative bimetallic mechanisms involving simultaneous superior to the monomeric cobalt(III) salen counterpart (96). activation of the nucleophile as well as Lewis acid activation Thus, the use of catalysts of type 97 allowed catalyst loadings of the oxetane. as little as 0.01 mol% to be used, with the rearranged products formed in up to 98% yield and 99% ee. The [68] enhanced reactivity of catalyst 97 (n = 1) can be explained Scheme 19. Ring expansion of oxetanes according to Lo and Fu. Cy = cyclohexyl. Fu, G.c. Tetrahedron 2001, 57, 2621.! by cooperative bimetallic mechanisms involving simultaneous activation of the nucleophile as well as Lewis acid activation 9060 www.angewandte.org  2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Angew. Chem. Int. Ed. 2010, 49, 9052 – 9067 of the oxetane.

Scheme 19. Ring expansion of oxetanes according to Lo and Fu.[68] Cy = cyclohexyl.

9060 www.angewandte.org  2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Angew. Chem. Int. Ed. 2010, 49, 9052 – 9067 C-H Functionalization with Oxetanes

§ C-H bonds α to an oxygen make possible selective radical-mediated transformations § Oxetanyl radical at α-carbon decreases puckering of 4-membered ring – BDE = 92.6 kcal/mol – Cyclobutane = 97.1 kcal/mol

Huie, R.; Kafafi, S. J. Phys. Chem. 1991, 95, 9340.!

§ C-H Functionalization approach is favorable because acid/base alkylations will not work

HO BEt3 (6 equiv.), air, r.t. O OMe MeO CHO O

77% Nagato, H. J. Org. Chem. 2005, 70, 2342.!

EtO2C Co(OAc)2 (0.1 mol%) CO2Et N-Hydroxyphthalimide (10 mol%) + EtO2C O CO2Et OH PhCN, r.t., 20 h O 76% Ishii, Y. Tet. Lett. 2002, 43, 3617.! COMMUNICATIONS

DOI: 10.1002/adsc.201400027 Photocatalytic Synthesis of Oxetane Derivatives by Selective C H Activation À Davide Ravelli,a,* Matteo Zoccolillo,a Mariella Mella,a and Maurizio Fagnonia a PhotoGreen Lab, Department of Chemistry, University of Pavia, Viale Taramelli 12, 27100 Pavia, Italy Fax: (+39)-0382-987323; phone:C-H (+ 39)-0382-987198; Activation e-mail: [email protected] with Oxetanes

Received: January 9, 2014; Revised: April 2, 2014; Published online: June 20, 2014 § C-H activation on could be intriguing for late stage functionalization of natural products Supporting information for this article is available on the WWW under http://dx.doi.org/10.1002/adsc.201400027. and drug targets containing oxetanes

Abstract: The selective C(ACHTUNGRE sp3) H activation at posi- À tion 2 in oxetanes has been accomplished by deca- tungstate photocatalysis under mild conditions. The resulting a-oxy radicals were trapped by electron- poor olefins resulting in the smooth preparation of 2-substituted oxetanes. The chemoselectivity in hy- drogen abstraction in substituted oxetanes contain- ing other H donating groups, such as CH2OH, CH2OAc and CHO, has been demonstrated in in- tramolecular models. Scheme 1. Known strategies (left side) and the approach Keywords: C H activation; decatungstate anion; proposed in this work (right side) for the preparation of oxe- À Michael addition; oxetanes; photocatalysis§ Decatungstate aniontane derivatives.photocatalysis has been done on THF ehters

– C2-H bond BDE = 92.6 kcal/mol vs THF = 92.8 kcal/mol oxetanes have been more rarely prepared from a pre- formed four-membered heterocyclic ring, which would involve the direct derivatization of oxetanesO by O [(n-Bu) N] [W O ] (2 mol%) The oxetane ring is present in important natural prod- +C H activation (path4 d).4 This10 32 is not possible under À SO2Ph 310 nm, MeCN, 8 h ucts and in drugs, as in the case of Taxol (Paclitaxel), acidic or basic conditions, since these may induce SO2Ph a widely known antitumoral.[1] The introduction of competitive ring opening,[4]76%the only known exception this moiety in bioactive compounds has a remarkable being that of 2-phenyloxetane, where the phenyl effect in terms of solubility and stiffening of the struc- group was able to confer a sufficient stabilization to [1–3] [9] ture, as well as metabolic stability. With regard to the carbanion formed. On the other hand,Fagnoni carbene, M. Adv. Synth. Catal. 2014, 356, 2781.! the biological effect, the oxetane ring has been con- insertion is unselective and ring enlargement or ring sidered as a surrogate of carbonyl groups, avoiding opening usually accompany this process making this the high reactivity of the latter, and of a gem-dimethyl approach unpractical.[10] Activation of the C H bonds À group, avoiding an increase in lipophilicity, and fur- by using a radical initiator and trapping of the result- ther has a higher Lewis basicity with respect to other ing a-oxy radical by an electron-poor olefin would alicyclic ethers.[1,4,5] Recently, oxetanes attracted con- seem a more appealing approach, but has been siderable interest in the field of polymer chemistry, reported only in a few cases,[11] starting from parent [11c] particularly for their application in opto-electronic de- oxetane by activation via Et3B/O2, [6] ACHTUNGRE [11d] [11a] vices. This makes it desirable that methods for intro- NHPI/Co(OAc)2, or benzoyl peroxide (but only ducing this moiety into pre-existing scaffolds be avail- with the oxetane as solvent). Furthermore, this ap- able, particularly for aliphatic derivatives. proach is discouraged by the high activation energy Of the main routes for the preparation of oxetane for H abstraction from oxetanes measured with the derivatives, the intramolecular Williamson reaction[1b] sulfate radical anion,[12] although the BDE (bond dis- (Scheme 1, path a) and the Paternò–Büchi photo- sociation energy) for the C2 H bond in oxetane [1b,7] 1 À chemical [2+2]cycloaddition (path b) involve (92.6 kcalmolÀ ) was calculated to be very similar to 1 [12] building of the four-membered ring and the third one, that of THF (92.8 kcalmolÀ ). treatment of epoxides with dimethylsulfoxonium The mild and effective generation of a-oxy radicals methylide (path c), ring enlargement.[8] C-Substituted via photocatalysis by the decatungstate anion[13] {used

Adv. Synth. Catal. 2014, 356, 2781 – 2786  2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim 2781 Davide Ravelli et al. COMMUNICATIONS

Table 1. Photocatalytic functionalization of oxetane 1a.[a]

Entry [1a](xM) Solvent Irradiation Conditions 2a Consumption% (Time, [h]) 3 Yield [%][b] 1 0.1 MeCN 310 nm 85 (24) 31[c] 2 0.3 MeCN 310 nm 82 (4) 74 100 (8) 76 100 (24) 65 3[d] 0.3 MeCN 310 nm 100 (8) 61

4 0.3 MeCN-H2O (5/1) 310 nm 100 (8) 65 [e] 5 0.3 MeCOMe-H2O (5/1) 310 nm 100 (8) 37 6 0.3 MeCN 366 nm 80 (8) 74 7 0.3 MeCN SolarBox 100 (8) 66 8[f] 0.3 MeCN 310 nm 61 (24) 8 9[g] 0.3 MeCN – 2 (24) 0 [a] ACHTUNGRE Reaction conditions: 1a (3x mmol), 2a (0.3 mmol), [(n-Bu)4N]4[W10O32] (TBADT, 2 mol%) in 3 mL of solvent. [b] Gas chromatography (GC) yields referred to the consumption of the limiting reagent (2a) using n-undecane as internal standard. [c] Some filming on the tube walls has been observed. [d] 1 mol% TBADT used. [e] The formation of by-products has been observed by GC analysis. [f] No TBADT used. [g] Reaction carried out in the absence of light.

ACHTUNGRE as tetrabutylammonium salt, [(n-Bu)4N]4[W10O32], lated conditions, entries 6 and 7). Blank experiments TBADT} was recently reported, as well as the ensuing demonstrated that both the photocatalyst and light application to C H bond activation in cyclic ethers, are required for the success of the reaction (entries 8 À such as tetrahydrofuran (THF),[14] 1,3-benzodiox- and 9). oles[15] and crown ethers.[16] Indeed, we surmised that The mechanism is outlined in Scheme 2. Hydrogen photocatalysis by the decatungstate anion could offer abstraction from position 2 by the excited photocata- an alternative entry to substituted oxetanes lyst generatesC-H theActivation corresponding oxetan-2-ylwith Oxetanes radical. (Scheme 1, path d). This is in turn trapped by the electron-poor olefin to The photocatalyzed reaction of oxetane (1a) in the give a radical adduct, that regenerates the reduced O O [(n-Bu) N] [W O ] (2 mol%) R R + 4 4 10 32 R presence of phenyl vinyl sulfone (2a) was first ex- photocatalyst and affords the desired product. EWG 310 nm, MeCN R EWG plored (Table 1). Thus, an acetonitrile solution con- An alternative§ Reaction activation requires mechanism both light and would involve taining an equimolar amount (0.1M) of 1a and 2a was an initial electronphotocatalyst transfer from oxetane 1 to excitedOxetane Olefin Product Yield (%) irradiated in a multilamp apparatus fitted with 10 decatungstate to give the corresponding radical cationO O – Requires electron deficient phosphor-coated lamps (15 W each; irradiation cen- CN 70 olefins CN tered at 310 nm) for 24 h in the presence of TBADT O 3 O (2”10À M; 2 mol%). Encouragingly, the hoped for CO2Me 60 CO2Me product 2-[2-(phenylsulfonyl)ethyl]oxetane (3) was O formed in 31% yield (85% sulfone consumption; O O O entry 1 in Table 1). Increasing the amount of 1a to 62

3 equiv., allowed us to reach complete consumption NC CN after 8 h irradiation and led to a higher yield (76%) O O CN of 3 (prolonged irradiation, 24 h, caused partial degra- CN 64 dation, though, see entry 2). Halving the amount of O O O the photocatalyst (1 mol%) diminished the yield 50 (entry 3), as did passing to different reaction media O

(acetonitrile/water or acetone/water mixtures, in the O O latter case by-products were formed, entries 4 and 5). SO2Ph 52 SO Ph On the contrary, the yield of sulfone 3 was maintained 2

O O CO2Me under different irradiation conditions (phosphor- Scheme 2. Proposed mechanism for the TBADT-photocata- CO2Me 69 CO2Me coated lamps centered at 366 nm or solar-light simu- lyzed functionalization of oxetanes. CO2Me

Fagnoni, M. Adv. Synth. Catal. 2014, 356, 2781.! 2782 asc.wiley-vch.de  2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Adv. Synth. Catal. 2014, 356, 2781 – 2786 C-H Activation with Oxetanes

§ Selectivity slightly favors methine activationDavide Ravelli et al. COMMUNICATIONS – H-abstraction is reversible, sterics hinder coupling in formation of tertiary radical

Scheme 3.

more hindered 2h leaves room to back-hydrogen Fagnoni, M. Adv. Synth. Catal. 2014, 356, 2781.! transfer thus causing the preferred addition of the (less hindered) secondary radical (from the methylene position). In accord with this hypothesis, the reaction with 2h is slower. The diastereoselectivity observed in 17a is probably related to a minimization of steric re- pulsion at the radical-adduct level. With the aim of further investigating the effect of the substitution pattern on the process, we purposely synthesized 2-tert-butyloxetane (1d), bearing a bulky group at the 2-position. Indeed, in the reaction with Scheme 4. acrylonitrile 2b (Scheme 3b), only two isomers were formed, viz. 18a and 18b resulting from the activation of the methylene and the methine hydrogens, respec- tively, although in a modest overall yield (35%). Ac- hol had previously been found to be not susceptible tually, the ratio 18a/18b=1.3 (statistically corrected) to TBADT photocatalysis.[15] However, acetylation of indicates that the presence of a bulky group disfavors the hydroxy group inhibited the competing path and the abstraction of the hydrogen from the carbon atom restored the reaction at the oxetane moiety. Thus 1f bearing the substituent, probably due to steric reasons reacted with 2h to give compound 20 in 52% yield (compare the 16a/16b ratio above). Analogously, (Scheme 4), noteworthy in a completely diastereose- a more sterically demanding substituent drives to the lective fashion, probably through a path again guided formation of a single diastereoisomer 18a (trans con- by steric hindering at the radical-adduct stage. figuration). In the search of extending the functionalization to The above experiments established the selective hy- position 3, a further intramolecular model was pre- drogen abstraction by TBADT from position 2 of ox- pared, oxetanecarbaldehyde 1g (Scheme 5). On the etane. Competition with other potential hydrogen basis of our experience in photocatalyzed C H activa- À donors was next tested by intramolecular models. tion, we expected a preferential activation of the Commercially available (3-methyloxetan-3-yl)metha- C(ACHTUNGRE sp2) H hydrogen.[19] Gratifyingly, a fully chemose- À nol (1e; again 5 equiv. used) gave no photocatalyzed lective activation of the formyl hydrogen led in reaction in the presence of isopropylidenemalononi- a good overall yield to oxetanes 21 and 22 from trile 2h (Scheme 4). Apparently, H abstraction from olefin 2a,[20] and to oxetane 23 as the exclusive prod- the CH OH group competed and the resulting a- uct from 2g (Scheme 5 and Scheme 6). The high reac- À 2À oxy radical was converted to CHO +H+ by oxida- tivity of the CHO group allowed for the use of À À tion by the ground state photocatalyst preventing any a lower amount of the oxetane (0.15M rather than addition to the olefin.[14] Analogously, piperonyl alco- 0.3M). The acyl radical formed decarbonylated

2784 asc.wiley-vch.de  2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Adv. Synth. Catal. 2014, 356, 2781 – 2786 Davide Ravelli et al. COMMUNICATIONS

C-H Activation with Oxetanes

§ Competetion between 2-position of oxetanesCOMMUNICATIONS and other potential hydrogenPhotocatalytic donors Synthesis of Oxetane Derivatives

Experimental Section Scheme 3. Typical Procedure for the Photocatalyzed Functionalization of Oxetanes more hindered 2h leaves room to back-hydrogen A solution (30 mL) of the oxetane derivative (1, 0.15–0.5M; 4.5–15 mmol) and the olefin (2, 0.1M; 3 mmol) in the pres- transfer thus causing the preferred addition of the 3 (less hindered) secondary radical (from the methylene ence of 200 mg of TBADT (2”10À M; 0.06 mmol) in MeCN was poured into two quartz tubes and purged for 5 min with position). In accord with this hypothesis, the reaction nitrogen, the tubes were then septum capped and irradiated with 2h is slower. The diastereoselectivity observed in for the required time (8–40 h; see Table 2) with ten 15 W 17a is probably related to a minimization of steric re- phosphor-coated lamps (emission centered at 310 nm, Scheme 5. pulsion at the radical-adduct level. except for the reaction between 1g and 2a, where emission With the aim of further investigating the effect of centered at 366 nm was employed). The solvent was re- the substitution pattern on the process, we purposely moved under vacuum from the photolyzed solution and the synthesized 2-tert-butyloxetane (1d), bearing a bulky product isolated by purification of the residue by column chromatography (hexane/ethyl acetate as eluants). group at the 2-position. Indeed, in the reaction with Scheme 4. acrylonitrile 2b (Scheme 3b), only two isomers were formed, viz. 18a and 18b resulting from the activation § R=H, product formed was aldehyde of the methylene and the methine hydrogens, respec- § Aceylation inhibited competing pathway tively, although in a modest overall yield (35%). Ac- hol had previously been found to be not susceptible Acknowledgements tually, the ratio 18a/18b=1.3 (statistically corrected) to TBADT photocatalysis.[15] However, acetylation of We thank Prof. Angelo Albini (University of Pavia) for fruit- indicates that the presence of a bulky group disfavors the hydroxy group inhibited the competing path and ful discussion and Andrea Gandini (University of Pavia) for the abstraction of the hydrogen from the carbon atom restored the reaction at the oxetane moiety. Thus 1f Scheme 6. his assistance. bearing the substituent, probably due to steric reasons reacted with 2h to give compound 20 in 52% yield Fagnoni, M. Adv. Synth. Catal. 2014, 356, 2781.! (compare the 16a/16b ratio above). Analogously, (Scheme 4), noteworthy in a completely diastereose- a more sterically demanding substituent drives to the lective fashion, probably through a path again guided References formation of a single diastereoisomer 18a (trans con- by steric hindering at the radical-adduct stage. easily[21] and a mixture of alkylated and acylated oxe- figuration). In the search of extending the functionalization to tanes was formed with a good trap (e.g., 2a),[22] [1] a) J. A. Burkhard, G. Wuitschik, M. Rogers-Evans, K. The above experiments established the selective hy- position 3, a further intramolecular model was pre- whereas exclusive alkylation resulted with 2g Müller, E. M. Carreira, Angew. Chem. 2010, 122, 9236– drogen abstraction by TBADT from position 2 of ox- pared, oxetanecarbaldehyde 1g (Scheme 5). On the (Scheme 6). The presence of the formyl group al- 9251; Angew. Chem. Int. Ed. 2010, 49, 9052–9067; b) G. etane. Competition with other potential hydrogen basis of our experience in photocatalyzed C H activa- lowed for the formation of a C C bond at the ex- Wuitschik, E. M. Carreira, B. Wagner, H. Fischer, I. À À Parrilla, F. Schuler, M. Rogers-Evans, K. Müller, J. donors was next tested by intramolecular models. tion, we expected a preferential activation of the pense of another C C bond. Commercially available (3-methyloxetan-3-yl)metha- C(ACHTUNGRE sp2) H hydrogen.[19] Gratifyingly, a fully chemose- À Med. Chem. 2010, 53, 3227–3246. À In conclusion, the alkylation of oxetanes (in an [2] M. Wang, B. Cornett, J. Nettles, D. C. Liotta, J. P. nol (1e; again 5 equiv. used) gave no photocatalyzed lective activation of the formyl hydrogen led in amount from 1.5 to 5 equivalents for the most hin- reaction in the presence of isopropylidenemalononi- a good overall yield to oxetanes 21 and 22 from Snyder, J. Org. Chem. 2000, 65, 1059–1068. dered terms with respect to the olefin) by photocata- [3] a) J. A. Burkhard, G. Wuitschik, J.-M. Plancher, M. trile 2h (Scheme 4). Apparently, H abstraction from olefin 2a,[20] and to oxetane 23 as the exclusive prod- lytic activation appears to be a large-scope synthetic Rogers-Evans, E. M. Carreira, Org. Lett. 2013, 15, the CH OH group competed and the resulting a- uct from 2g (Scheme 5 and Scheme 6). The high reac- À 2À method. This is particularly suited for inserting this 4312–4315; b) A. F. Stepan, K. Karki, W. S. McDonald, oxy radical was converted to CHO +H+ by oxida- tivity of the CHO group allowed for the use of À À group into electrophilic alkenes and for the prepara- P. H. Dorff, J. K. Dutra, K. J. DiRico, A. Won, C. Sub- tion by the ground state photocatalyst preventing any a lower amount of the oxetane (0.15M rather than ramanyam, I. V. Efremov, C. J. ODonnell, C. E. Nolan, [14] tion of aliphatic derivatives, significantly supplement- addition to the olefin. Analogously, piperonyl alco- 0.3M). The acyl radical formed decarbonylated ing established procedures for the access to this in- S. L. Becker, L. R. Pustilnik, B. Sneed, H. Sun, Y. Lu, creasingly used class of compounds. Indeed, our ap- A. E. Robshaw, D. Riddell, T. J. OSullivan, E. Sibley, S. Capetta, K. Atchison, A. J. Hallgren, E. Miller, A. 2784 asc.wiley-vch.de  2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Adv. Synth. Catal. 2014, 356, 2781 – 2786 proach is the first case reported of the photochemical Wood, R. S. Obach, J. Med. Chem. 2011, 54, 7772–7783. generation of a-oxy radicals from oxetanes. Notice [4] H. C. Hailes, J. M. Behrendt, Oxetanes and Oxetenes: for example that none of the products formed (3–23) Monocyclic, in: Comprehensive Heterocyclic Chemistry was previously known. In particular, no evidence for III, Vol. 2, Ch. 2.05, (Ed.: A. R. Katritzky), Pergamon, their synthesis by the otherwise convenient Paternò– Oxford, 2008, pp 321–364. Büchi reaction was found in the literature (which [5] G. Wuitschik, Oxetanes in Drug Discovery, PhD Disser- would involve a photoreaction between formaldehyde tation, Zürich, 2008. and a mono-substituted olefin or between ethylene [6] a) H. Kudo, T. Nishikubo, J. Polym. Sci. Part A J. and an aldehyde). Furthermore, an indirect method to Polym. Sci. Part A: Polym. Chem. 2007, 45, 709–726; introduce the oxetane moiety has been devised by ac- b) A. Charas, J. Morgado, Curr. Phys. Chem. 2012, 2, tivation of a formyl group. The reaction conditions 241–264. [7] a) M. Abe, in: Handbook of Synthetic Photochemistry, are very mild and solar light could be used to pro- [23] (Eds.: A. Albini, M. Fagnoni), Wiley-VCH, Weinheim, mote the process. 2010, pp 217–239; b) A. G. Griesbeck, S. Bondock, in: CRC Handbook of Organic Photochemistry and Photo- biology, 2nd edn., (Eds.: W. Horspool, F. Lenci), CRC Press, Boca Raton, 2004, pp 60–1/60–21; c) M. D’Auria,

Adv. Synth. Catal. 2014, 356, 2781 – 2786  2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim asc.wiley-vch.de 2785 Wrap Up

§ The use of oxetane containing molecules for drug discovery will likely increase greatly in the coming years • Either from chemists push to use this new exciting molecule • Or from physiochemical changes induced by replacing previous functionalities

§ The added polarity effect of oxetane incorporation and lack of reactivity could be utilized widely for complex molecule synthesis and polymer formation for solubility increase

§ Transition metal reactions with oxetanes have mainly focused on carbonylative reactions

§ Only now are people beginning to use oxetanes as substrates for new, exciting transition metal catalyzed reactions Questions

1. Predict the product for the following reactions.

TMSCN O Et2AlCl r.t., 24 h

O Ph nBuLi, BF3-OEt2 3. Please propose a mechanism for the reaction THF, -78 °C below. O MeO2C CO2Me TFA O O Me Ph NH 10 mol% [Ni(cod)2] 20 mol% iPr CO Me O + 2 OtBu O toluene, 0 °C, 8 h O CO2Me Me Me O Me h! Ph H O

N Me

Ph 2. Please propose a mechanism for the reaction below.

OTIPS O TIPSO R O O HNTf2 (10 mol%) O DCM, r.t. H R Question 1

TMSCN O Et2AlCl r.t., 24 h NC OTMS

O Ph OH nBuLi, BF -OEt 3 2 Ph THF, -78 °C

O OH Ph TFA Ph NH NH O O OtBu O

O O O h! O Ph N Ph H Me N Me Ph

Ph Question 2

!"#$%&'! !"#$%&'(&")*"+$,-'$."/*012$(3/$4(')"25*3)$4(6&-3"$%73&8"+*+ ()*

+ O H OH H O O C O+ O – NTf2 R TIPS + O TIPS O 16 R 17 H+ 15 •• path a OTIPS OH R OTIPS TIPSNTf HNTf2 2 O O O O O O 12 R R 14 TIPSNTf2 18 path b R OH [2+2] O + H+ H OTIPS + 13 O O

O •• O TIPS ••O R R – O TIPS O TIPS NHTf 19 21 2 R 20 !89:/:;- $e=(0+*Q="$1"68(3*+1

O O '*3)2"VK(3+*-3$K'-&-6-=$6(3$(=+-$Q"$(KK=*"/$&-$&8"$+73&8"2 +*+$-,$=(6&-3"+$#*&8$-&8"'$'*3)$+*W"+$R+*0D1SO$G-'$"V(1K="A Ar +&('&*3)$,'-1$&8"$,*N"2A$ +"N"32A$"*)8&2A$ (3/$XM21"1Q"'/ O O '*3)$(6"&(=+A$&8"$6-''"+K-3/*3)$+"N"32A$3*3"2A$&"32A$(3/$XY2 12 22 1"1Q"'"/$=(6&-3"+$6(3$(==$Q"$,-'1"/$*3$)--/$&-$"V6"=="3& <5% yield 7*"=/+O$Z3$&"'1+$-,$&8"$(=@"3"$6-3,*)0'(&*-3$*3$&8"$=(6&-3" !89:/:;* $I,,"6&$-,$&8"$=*3@"' K'-/06&+A$ +"N"32$ &-$ 3*3"21"1Q"'"/$ =(6&-3"+$ #"'"$ -Q2 &(*3"/$(+$($+*3)="$)2(=@"3"A$(3/$&"321"1Q"'"/$=(6&-3"$+*2 #(+$-Q&(*3"/$(+$($X[X$1*V&0'"$-,$)\*$*+-1"'+O$Z3$6-3&'(+&A a R R -3=7$&8"$*$*+-1"'$#(+$-Q+"'N"/$,-'$&8"$=(')"2'*3)$=(6&-3" b +*1O b n+2 O OTIPS O O ;8"$=(6&-3"$,-'1(&*-3$'"(6&*-3$6-0=/$Q"$,0'&8"'$0&*=*W"/ 24 25 ,-'$*&"'(&*N"$'*3)$"VK(3+*-3+$R%68"1"$YSO$]3/"'$&8"$+&(32 /('/$6-3/*&*-3+A$&8"$'"(6&*-3$-,$&8"$,*N"21"1Q"'"/$676=*6 (6"&(=$+30$(3/$(=@73"$,(.$(,,-'/"/$+"N"321"1Q"'"/$=(62 a [2+2] &-3"$+*0A$#8*68$#(+$"(+*=7$6-3N"'&"/$&-$(6"&(=$+4$(,&"' 87/'-)"3(&*-3$(3/$'"/06&*N"$(67=(&*-3O$%0Q^"6&*-3$-,$(6"2 R &(=$ +4$ &-$ ($ +"6-3/$ '*3)$ "VK(3+*-3$ #*&8$ (=@73"$ ,(5$,0'2 3*+8"/$ 3*3"21"1Q"'"/$ =(6&-3"$ +*6O$ ;8"$ *&"'(&*N"$ '*3)2 n O + "VK(3+*-3$+"U0"36"$6-0=/$Q"$6-3+*/"'"/$($,-'1(=$_`aJaJb 23 TIPSO 676=*W(&*-3$'"(6&*-3O$Z3$K'*36*K="A$&8"$*&"'(&*N"$'*3)2"VK(32 13 +*-3$K'-&-6-=$6-0=/$Q"$,0'&8"'$(KK=*"/$,-'$1-'"$&*1"+$&- !89:/:;3 $f"+*)3$-,$($3"#$'*3)$"VK(3+*-3$+&'(&")7 (++"1Q="$(==$+*W"+$-,$=(6&-3"+$,'-1$+1(==$676=*6$(6"&(=+O

;8"$K'-K-+"/$1"68(3*+1$*+$/"K*6&"/$*3$%68"1"$cO$;8"$'"2 Downloaded by: IP-Proxy University of Texas at Austin, Austin Libraries. Copyrighted material. 60'+-'$ (3/$ 4"#*+$ (6*/$ FGB>HI&J$(+$&8"$K'-1-&"'$#*&8 (6&*-3$Q")*3+$#*&8$FGB2K'-1-&"/$"=*1*3(&*-3$-,$(=@-V*/" JALAM26-==*/*3"$(+$&8"$(//*&*N"O$P+$+8-#3$*3$%68"1"$EA$( ,'-1$(6"&(=$ +3$&-$,-'1$-V-6('Q"3*01$+4O$Z3$ &8"$1"(32 '(3)"$-,$"*)8&21"1Q"'"/$=(6&-3"+$#*&8$/*,,"'"3&$+0Q+&*&02 #8*="A$+*=-V7$(=@73"$,($*+$(6&*N(&"/$&-$,-'1$73-=(&"$+7A "3&+$(&$/*,,"'"3&$K-+*&*-3+$6(3$Q"$-Q&(*3"/$#*&8$)--/$&- #8*68$&8"3$03/"')-"+$_JaJb$676=-(//*&*-3$#*&8$-V-6('Q"2 8*)8$ ",,*6*"367$ R+*.D/SO$ P$ /*N"'+"$ +"&$ -,$ ,036&*-3(= 3*01$+4$&-$,0'3*+8$&8"$@"7$-V"&"3*01$*3&"'1"/*(&"$()O$G*2 )'-0K+$('"$6-1K(&*Q="$#*&8$&8"$'"(6&*-3$6-3/*&*-3+O$T-12 3(==7A$ "="6&'-676=*6$ '*3)$ -K"3*3)$ -,$ &8"$ -V"&"3*01$ '*3) K('*3)$#*&8$&8"$K'"N*-0+$+&'(&")7A$&8"$K'"+"3&$1"&8-/$/-"+ /"=*N"'+$&8"$-Q+"'N"/$'*3)2"VK(3+*-3$K'-/06&$+*O$d"$K'-2 3-&$'"U0*'"$(3$('7=$)'-0K$,0+"/$*3$&8"$=(6&-3"$+@"="&-3A K-+"$&8(&$&8"$'-="$-,$(//*&*N"$JALAM26-==*/*3"$*+$&-$+&(Q*=*W" #8*68$ +*)3*,*6(3&=7$ "VK(3/+$ &8"$ +6-K"O$ .-'"-N"'A$ &8" &8"$-V-6('Q"3*01$+4$Q7$'"N"'+*Q=7$,-'1*3)$&8"$K7'*/*3*2

9:"-')$;8*"1"$<"'=()$$%&0&&)('&$>$!"#$?-'@ !"#$%&&$+),-A$'(A$BCBDBCE Angewandte. Communications

A proposed mechanism for the cycloaddition reaction is shown in Scheme 2. Typically, such a cycloaddition would begin with oxidative coupling between an alkyne and the Questioncarbonyl 3 group. [11,22,23] In this case, oxidative coupling would

Scheme 1. Products derived from the Ni-catalyzed cycloaddition of diynes 1b–1k and azetidinone 2d. Reaction conditions: azetidinone

2d(1 equiv), diyne 1b–1k (1.2 equiv), [Ni(cod)2] (10 mol%), IPr (20 mol%), toluene, 8h. products (3jd and 3kd, Scheme 1), which contain a spirocyclic backbone, were obtained in good yield. 3-oxacyclobutanone (2e) and diyne 1acould also undergo the Ni/IPr-catalyzed cycloaddition reaction using the standard reaction conditions, thus giving dihydrooxocine 3ae in good yield [Eq. (2)].

Scheme 2. Proposed mechanism of the cycloaddition reaction.

afford a spirocyclic intermediate M1. Subsequent insertion of

To address the question of regioselectivity when using an the pendant alkyne would give the intermediate M2, which unsymmetrical diyne, diyne 1l, which contains a terminal could then undergo either reductive elimination (to afford alkyne and a methyl-substituted alkyne, was subjected to the products such as 3nd and 3od) or b-carbon elimination to optimized reaction conditions; only one regioisomer was afford metallacycle M . Finally, C C bond forming reductive 3 À observed [3ld; Eq. (3)]. Another unsymmetrical diyne, 1m, elimination occurs to give the dihydroazocine product and the the termini of which differ electronically rather than sterically, catalyst. In the cycloaddition of unsymmetrical diyne 1l, the was also evaluated; again, only one regioisomer was observed regioselectivity is governed by the difference in the size of the [3md, Eq. (3)]. substituents on the alkynes. That is, oxidative coupling Interestingly, when diynes 1n and 1o, which were between the alkyne bearing the larger group occurs first to expected to afford six-membered-ring-fused dihydroazocine minimize steric interactions between the group and the products, were evaluated, the expected products arising from ligand. Indeed, this type of regioselectivity has been found

C 2 C 3 bond cleavage were not obtained. Instead, spirocy- in various cycloaddition reactions catalyzed by nickel com- sp À sp clic pyran products were formed in good yields [3nd and 3od, plexes. However, in the cycloaddition of unsymmetrical diyne Eq. (4)].[16] 1m, the electronic nature of the alkyne, rather than steric

8604 www.angewandte.org  2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Angew. Chem. Int. Ed. 2012, 51, 8602 –8606