Abstract
The site-selective and chemoselective functionalization of alcohols in complex polyols remains a formidable synthetic challenge. Whereas significant advancements have been made in selective derivatization at the oxygen center, chemoselective oxidation to the corresponding carbonyls is less developed. In cyclic systems, whereas the selective oxidation of axial alcohols is well known, a complementary equatorial selective process has not yet been reported. Herein we report the utility of nitrogen-ligated (bis)cationic λ3-iodanes (N-HVIs) for alcohol oxidation and their unprecedented levels of selectivity for the oxidation of equatorial over axial alcohols. The conditions are mild, and the simple pyridine-ligated reagent (Py-HVI) is readily synthesized from commercial PhI(OAc)2 and can be either isolated or generated in situ. Conformational selectivity is demonstrated in both flexible 1,2-substituted cyclohexanols and rigid polyol scaffolds, providing chemists with a novel tool for chemoselective oxidation.
Graphical Abstract
The chemoselective and site-selective modification of complex molecular scaffolds persists as a formidable synthetic challenge. One area of particular interest is the selective reaction of one hydroxyl group among numerous sterically similar sites on a polyol scaffold. Whereas significant advancements have been made toward derivatization at oxygen,1 selective reaction at the α-carbon through oxidation to the corresponding carbonyls is less developed. In a significant advancement, the Hartwig laboratory recently disclosed an elegant approach to the selective oxidation of hindered secondary alcohols in the presence of primary sites via rhodium-catalyzed transfer hydrogenation.2 In cyclic systems, conformational effects can drive selectivity; Cr(VI) oxidants such as PCC will selectively oxidize axial alcohols due to steric relief driving the breakdown of the intermediate chromate ester.3,4 Whereas equatorial alcohols are more sterically accessible,5 to the best of our knowledge, a broadly selective method for the preferential oxidation of equatorial over axial alcohols has not been reported. Thus, the development of such a method would provide chemists with a powerful new tool in the chemoselective oxidation of complex molecules.
Recently, our laboratory has reported on the synthetic applications of (bis)cationic nitrogen-ligated λ3-iodanes [PhI-(het)2]2OTf−, or N-HVIs (1) (Scheme 1).6–9 Whereas N-HVIs have been known for over 20 years, reports on the synthetic utility of these reagents are scarce.6,10,11 N-HVIs possess two datively bound heterocyclic nitrogen ligands on the central iodine, resulting in significantly enhanced oxidation potentials relative to traditional λ3-iodanes and opening up new modes of reactivity to this reagent class.11,12 Adding to their appeal, N-HVIs are readily synthesized from commercially available PhI(OAc)2 using just TMSOTf and the desired heterocycle (Scheme 1A), leading to free-flowing white solids.6
Scheme 1.
N-HVIs: Synthesis and Reactivity
During our recent efforts to develop an electrophilic ring expansion of secondary alcohols with N-HVIs (2 → 3, Scheme 1B),9 we observed that the relative alcohol stereochemistry had a significant effect on the reaction pathway. In using model substrate 2-Me-cyclohexanol (2), cis-2 gave a clean rearrangement to cyclic ether 3, whereas trans-2 led to significant amounts of competitive ketone formation (4). We rationalized this divergent reactivity through differences in the conformational equilibria of the two substrates. The trans-2 placed the alcohol equatorial in the major conformer, resulting in an intermediate iodate ester that could readily achieve the necessary antiperiplanar orbital alignment with the α-hydrogen (2-trans-OHeq), enabling facile oxidation. In contrast, the cis-2 was biased toward an axial alcohol conformer, wherein significant steric clash with the 1,3-diaxial hydrogens inhibited access to the necessary orbital alignment, thus leading to rearranged products (2-cis-OHax). On the basis of this model, we recognized that if the N-HVI reactivity could instead be tuned to favor oxidation, this could lead to the first general method for the chemoselective oxidation of equatorial over axial alcohols. Furthermore, this would provide a rare example of broad-scope alcohol oxidation with a λ3-iodane reagent as the sole oxidant,13 providing practical advantages over the current λ5-iodane oxidants.12,14,15
Herein we report the successful development of two practical protocols for alcohol oxidation using simple pyridine-ligated N-HVI, Py-HVI (5), and its utility in the chemoselective oxidation of equatorial alcohols. The oxidation can be performed either with isolated Py-HVI or via its in situ generation, enabling a one-pot procedure from commercial PhI(OAc)2 (Scheme 1C). The oxidation shows a broad substrate scope and functional group tolerance. In cyclic substrates, Py-HVI displays excellent selectivity for equatorial alcohols in both conformationally flexible 1,2-substituted cyclohexanols and rigid polyol scaffolds, providing chemists with a general approach to the chemoselective oxidation of equatorial alcohols for the first time.
To begin our study, efficient conditions for direct alcohol oxidation with N-HVIs needed to be established. Using 4-phenyl-2-butanol (6) as a model substrate, treatment with 2.0 equiv of Py-HVI in 1,2-dichloroethane (DCE) at room temperature gave a modest 37% yield of the corresponding ketone (7) (Table 1, entry 1). Increasing the temperature to 60 °C resulted in near-quantitative oxidation in just 2.5 h, producing 7 in 96% isolated yield (entry 2). Decreasing the equivalents of Py-HVI to either 1.0 or 1.5 (entries 3 and 4) also gave high conversion but with decreased efficiency, taking 8 and 12 h to reach completion, respectively, and thus 2.0 equiv of Py-HVI was used for the remainder of our studies. A brief solvent screen found that acetonitrile was equally effective (entry 5), obviating the need for a halogenated solvent if so desired. To gain more insight into the role of the nitrogen ligand, a small library of electronically and sterically diverse N-HVIs was screened. The use of ortho-substituted (8) or more electron-rich (9, 10) N-HVIs also gave excellent yields (entries 7–9), whereas the highly reactive p-CF3-Py-HVI (11) was not as effective (entry 10).
Table 1.
Oxidation with Isolated N–HVI
 | ||||
|---|---|---|---|---|
| entry | N-HVI | solvent | T (°C) | yield (%) |
| 1 | Py-HVI (5) | DCE | r.t. | 37a |
| 2 | Py-HVI (5) | DCE | 60 | 96 |
| 3 | Py-HVI (5)b | DCE | 60 | 84 (95)c |
| 4 | Py-HVI (5)d | DCE | 60 | 52 (84)e |
| 5 | Py-HVI (5) | ACN | 60 | 97 |
| 6 | Py-HVI (5) | THF | 60 | 45 |
| 7 | 2-MePy (8) | DCE | 60 | 97a |
| 8 | 4-OMe (9) | DCE | 60 | 89a |
| 9 | 4-NMe2 (10) | DCE | 60 | 75a |
| 10 | 4-CF3 (11) | DCE | 60 | 63a |
NMR yield with CH2Br2 as an internal standard
1.5 equiv of Py-HVI.
Yield after 8 h.
1.0 equiv of Py-HVI.
Yield after 12 h.
We then wished to further simplify our method by developing a one-pot protocol via the in situ generation of Py-HVI, avoiding the need for its isolation and storage. This would also benefit the broader application of N-HVIs because these reagents can be moisture-sensitive and prone to degradation upon prolonged storage.9 The initial application of our standard conditions for N-HVI synthesis using TMSOTf, followed by the addition of the substrate, resulted in variable yields of 7 along with significant amounts of silyl-protected 6. The use of the bulkier TBSOTf as the silyl activator suppressed alcohol protection, and a screen of hydrogen bonding and Brønsted acid additives8,9,16–18 found that the use of AcOH gave a consistently excellent yield of 7 (Scheme 2). (See the Supporting Information for additional optimization details.) Several control reactions were conducted to determine if Py-HVI was, in fact, the active oxidant under these in situ conditions (Scheme 2, inset). Neither PhI(OAc)2 nor PhI(O2CCF3)2 alone produced any oxidation product, and PhI(OAc)2/AcOH gave only trace oxidation, indicating that Brønsted acid activation alone was insufficient. Finally, the exclusion of pyridine from the reaction resulted in significant substrate decomposition, likely through the reaction of highly reactive [PhI(OAc)]OTf, formed upon the treatment of PhI(OAc)2 with R3SiOTf.7b,17
Scheme 2.
Oxidation with in-Situ-Generated Py-HVI
The scope of the oxidation was then examined using both in-situ-generated (condition A) and isolated Py-HVI (5) (condition B) and was found to be quite general (Scheme 3). A variety of acyclic, cyclic, and benzylic alcohols (6, 12–19) gave good to excellent yields. The excellent yield of cyclopropyl substrate 14 provides evidence that the reaction does not proceed through a radical pathway. More functionalized substrates, including those with halogens (20, 26), acetate (21), alkyne (28), and saturated heterocycles (22, 23), all proceeded in high yield. We were pleased to see that an oxidatively sensitive thiophene and boronic ester were both well tolerated (24, 25) as well as a pyridine substrate (27), which underwent smooth oxidation rather than ligand exchange with the N-HVI.
Scheme 3. Alcohol Oxidation Scopec.
aRun in ACN-d3. bNMR yield, CH2Br2 internal standard. No isolated yield obtained due to volatility of products. cMethod (A): PhI(OAc)2 (2.0 equiv), TBSOTf (4.0 equiv) pyridine (4.0 equiv), DCE, 10 min rt; then AcOH (1.0 equiv), ROH substrate, 3A MS, 60 °C. Method (B): Py-HVI (2.0 equiv), DCE, 60 °C.
With robust oxidation conditions in hand, we returned to the intriguing question of effecting equatorial-selective alcohol oxidation. Our scope studies revealed that the newly optimized conditions exhibited significantly enhanced reactivity for 1,2-trans-substituted cyclic substrates over the corresponding 1,2-cis isomers, in line with findings from our prior studies.6 We therefore chose to begin our selectivity studies by examining if this divergent reactivity could be leveraged to achieve selective oxidation between two conformationally flexible molecules. Thus a 1:1 mixture of cis- and trans-22 was treated with in-situ-generated Py-HVI under our standard conditions (Scheme 4). The reaction displayed exquisite levels of selectivity for the oxidation of trans-22; ketone 29 was produced in 46% yield, along with only 2% recovered trans-22 and the near-quantitative recovery of cis-22 (entry 1). In fact, the reaction of pure cis-22 produced no ketone products, even upon prolonged heating, eventually producing small amounts of nonspecific degradation products. To further establish the novelty of this cis/trans chemoselectivity, the same mixture was subjected to a panel of common oxidants including Swern oxidation, pyridinium chlorochromate (PCC), Dess–Martin periodinane, and PhI(OAc)2/TEMPO,13i wherein the λ3-iodane acts only as a co-oxidant. As expected, no cis/trans selectivity was observed with Swern or Dess–Martin periodinane. PhI(OAc)2/TEMPO and PCC, on the contrary, showed slight selectivity, however, favoring the consumption of cis-22, likely arising from cis-22 slightly favoring the more rapidly oxidized axial alcohol conformation.
Scheme 4. Chemoselective Oxidation of 1,2-trans versus 1,2-cis Cyclic Alcohols.
Percentages based on total starting mmol cis/trans-22. aIsolated yields. bNMR yields only. CH2Br2 internal standard. cPercentages after 4 to 5 h of reaction time. dSelectivity ratio based on the percentage remaining of trans/cis-22 starting from 50% maximum yield.
Next, we examined the oxidation of conformationally locked scaffolds, wherein the reactivity of equatorial versus axial alcohols could be clearly established. Three trans-decalinol model substrates, 30, 31, and 32, were selected to examine the effect of both alcohol conformation and two-substitution on the reaction rate (Scheme 5). As predicted, equatorial alcohol 30 gave near-quantitative oxidation in only 30 min, producing 33 in 91% isolated yield (Scheme 5A). In marked contrast, axial alcohol 31 reacted sluggishly, producing only trace ketone after 1 h and only reaching 78% conversion after prolonged heating (Scheme 5B). Interestingly, alcohol 32, wherein the equatorial alcohol was flanked by a syn angular methyl group, analogous to equatorial alcohols in the 1,2-cis-cyclic alcohols (see Scheme 4), also displayed low reactivity, giving only 34% yield at 4 h.
Scheme 5. Chemoselective Oxidation of Conformationally Locked Scaffolds.
aFollowing Method B (premade Py-HVI) in CDCl3. bNMR yields using CH2Br2 internal standard. cFollowing Method A (in situ Py-HVI) dIsolated yield.
Finally, to demonstrate this novel chemoselectivity in a complex setting, deoxycholic acid (35), possessing both an equatorial A-ring alcohol and an axial C-ring alcohol, was subjected to oxidation with isolated Py-HVI (Scheme 5, inset). As expected, we observed an exclusively A-ring ketone 36 in 75% yield, with no products arising from either selective C-ring oxidation or double oxidation, even in the presence of excess Py-HVI. (See the Supporting Information for details.)
Taken together, these results demonstrate the ability of N-HVIs to effect the chemoselective oxidation of equatorial over axial alcohols for the first time. Given our findings, we now present a more comprehensive predictive model for the application of N-HVIs in conformationally selective oxidation (Scheme 6). In cyclic alcohols, chair conformations in which the alcohol is oriented axial suffer significant steric clash with the 1,3-diaxial hydrogens in the necessary conformation (37), thereby significantly inhibiting oxidation. In contrast, equatorial alcohols will undergo facile oxidation because the reactive iodate ester conformation is readily accessible (38). An exception exists in equatorial cases wherein an axial syn substituent at the two-position (e.g., angular methyl) results in a significant syn-pentane interaction between the equatorial iodate ester and the two-substituent (39), again leading to low reactivity. This final data point explains the general lack of reactivity in 1,2-cis-cyclic alcohols, whereby no active conformation for oxidation is readily accessible.
Scheme 6.
Predictive Model for Conformational Selectivity in N-HVI-Mediated Oxidation
In conclusion, we report a facile approach to the alcohol oxidation with a pyridine-ligated (bis)cationic λ3-iodane, Py-HVI, which is readily synthesized from commercial PhI(OAc)2 and can be either presynthesized or generated in situ. The enhanced reactivity of N-HVI reagents enables a rare example of oxidation with the λ3-iodane, avoiding the cost and safety concerns associated with Cr(VI)- and I(V)-based oxidants. In addition to its operational simplicity and broad substrate scope, the method displays exquisite selectivity for the oxidation of equatorial over axial alcohols, providing a powerful new tool with highly predictable reactivity for the chemoselective derivatization of complex scaffolds. N-HVI-mediated oxidation offers a general solution to a long-standing selectivity challenge in organic synthesis and should find widespread utility in total synthesis, particularly in the derivatization of complex polyols.
Supplementary Material
ACKNOWLEDGMENTS
We are grateful to the National Institutes of Health (NIH R01 GM123098) and the National Science Foundation (NSF CAREER 1752244) for financial support of this work. We thank Dr. Charles DeBrosse (Temple University) for NMR spectroscopic assistance and Dr. Charles W. Ross, III, Director, Automated Synthesis and Characterization at University of Pennsylvania Chemistry, for providing high-resolution mass spectral data.
Footnotes
The authors declare no competing financial interest.
ASSOCIATED CONTENT
Supporting Information
The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.orglett.9b02018.
Experimental details, analytical data, and NMR spectra (PDF)
REFERENCES
- (1).(a) Lewis CA; Miller SJ Angew. Chem., Int. Ed 2006, 45, 5616–5619. [DOI] [PubMed] [Google Scholar]; (b) Sun X; Lee H; Lee S; Tan KL Nat. Chem 2013, 5, 790–795. [DOI] [PMC free article] [PubMed] [Google Scholar]; (c) Peddibhotla S; Dang Y; Liu JO; Romo DJ Am. Chem. Soc 2007, 129, 12222–12231. [DOI] [PubMed] [Google Scholar]
- (2).Hill CK; Hartwig JF Nat. Chem 2017, 9, 1213–1221. [DOI] [PMC free article] [PubMed] [Google Scholar]
- (3).Roček J; Westheimer FH; Eschenmoser A; Moldoványi L; Schreiber J Helv. Chim. Acta 1962, 45, 2554–2567. [Google Scholar]
- (4).Muller P; Perlberger J-C Helv. Chim. Acta 1976, 59, 2335–2343. [Google Scholar]
- (5).Barton DHR Experientia 1950, 6, 316–320. [DOI] [PubMed] [Google Scholar]
- (6).For a seminal report on N-HVIs, see: Weiss R; Seubert J Angew. Chem., Int. Ed. Engl 1994, 33, 891–893. [Google Scholar]
- (7).For structural and computational studies of N-HVIs, see:Pell TP; Couchman SA; Ibrahim S; Wilson DJD; Smith BJ; Barnard PJ; Dutton JL Inorg. Chem 2012, 51, 13034–1300.Aprile A; Iversen KJ; Wilson DJD; Dutton JL Inorg. Chem 2015, 54, 4934–4939.
- (8).Kelley BT; Walters JC; Wengryniuk SE Org. Lett 2016, 18, 1896–1899. [DOI] [PubMed] [Google Scholar]
- (9).Walters JC; Tierno AF; Dubin AH; Wengryniuk SE Eur. J. Org. Chem 2018, 2018, 1460–1464. [DOI] [PMC free article] [PubMed] [Google Scholar]
- (10).For synthetic applications of N-HVIs beyond our laboratory, see:De Mico A; Margarita R; Piancatelli G Gazz. Chim. Ital 1995, 215, 325.Zhdankin VV; Maydanovych O; Herschbach J; Bruno J; Matveeva ED; Zefirov NS J. Org. Chem 2003, 68, 1018–1023.Kniep F; Walter SM; Herdtweck E; Huber SM Chem. - Eur. J 2012, 18, 1306–1310.Yuan Z; Cheng R; Chen P; Liu G; Liang SH Angew. Chem., Int. Ed 2016, 55, 11882–11886.
- (11).For a recent review on the applications of N-HVIs, including as oxidants in high-valent transition-metal chemistry, see: Corbo R; Dutton JL Coord. Chem. Rev 2018, 375, 69–79. [Google Scholar]
- (12).Yoshimura A; Zhdankin VV Chem. Rev 2016, 116, 3328–3435. [DOI] [PubMed] [Google Scholar]
- (13).For examples of alcohol oxidations with λ3-iodanes, see: with PhIO:Takaya T; Enyo H; Imoto E Bull. Chem. Soc. Jpn 1968, 41, 1032.Tohma H; Takizawa S; Maegawa T; Kita Y Angew. Chem, Int. Ed 2000, 39, 1306–1308. With PhIO/RuCl2(PPh3)3:Muller P; Godoy J Tetrahedron Lett. 1981, 22, 2361–2364.Muller P; Godoy J Helv. Chim. Acta 1983, 66, 1790–1975. With PhIO/Yb(NO3)3:Yokoo T; Matsumoto K; Oshima K; Utimoto K Chem. Lett. 1993, 22, 571–572. With PhI(OAc)2:Varma RS; Dahiya R; Saini RK Tetrahedron Lett. 1997, 38, 7029–7032.Narasaka K; Morikawa A; Saigo K; Mukaiyama T Bull. Chem. Soc. Jpn 1977, 50, 2773–2776.Ley SV; Thomas AW; Finch HJ Chem. Soc., Perkin Trans. 1 1999, 1, 669–671. With PhI(OAc)2/TEMPO:De Mico A; Margarita R; Parlanti L; Vescovi A; Piancatelli GJ Org. Chem 1997, 62, 6974–6977. With PhI(O2CCF3)2:Spyroudis S; Varvoglis A Synthesis 1975, 1975, 445–447.
- (14).(a) Dess DB; Martin JC J. Org. Chem 1983, 48, 4155–4156. [Google Scholar]; (b) Dess DB; Martin JC J. Am. Chem. Soc 1991, 113, 7277–7287. [Google Scholar]
- (15).Frigerio M; Santagostino M; Sputore SJ Org. Chem 1999, 64, 4537–4538. [Google Scholar]
- (16).(a) Colomer I; Chamberlain AER; Haughey MB; Donohoe TJ Nat. Rev. Chem 2017, 1, 1–12. [Google Scholar]; (b) Colomer I; Batchelor-McAuley C; Odell B; Donohoe TJ; Compton RG J. Am. Chem. Soc 2016, 138, 8855–8861. [DOI] [PubMed] [Google Scholar]
- (17).Izquierdo S; Essafi S; del Rosal I; Vidossich P; Pleixats R; Vallribera A; Ujaque G; Lledós A; Shafir AJ Am. Chem. Soc 2016, 138, 12747–12750. [DOI] [PubMed] [Google Scholar]
- (18).Dohi T; Yamaoka N; Kita Y Tetrahedron 2010, 66, 5775–5785. [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.