Download PDF
Synlett 2015; 26(12): 1643-1648
DOI: 10.1055/s-0034-1380616
synpacts
© Georg Thieme Verlag Stuttgart · New York

Structural Revision of Uprolide G Acetate: Effective Interplay between NMR Data Analysis and Chemical Synthesis

Liangyu Zhu
Department of Chemistry, The Hong Kong University of Science and Technology, Clear Water Bay, Kowloon, Hong Kong, P. R. of China   Email: rtong@ust.hk
,
Rongbiao Tong*
Department of Chemistry, The Hong Kong University of Science and Technology, Clear Water Bay, Kowloon, Hong Kong, P. R. of China   Email: rtong@ust.hk
› Author Affiliations
Further Information

Publication History

Received: 13 March 2015

Accepted after revision: 27 March 2015

Publication Date:
04 May 2015 (online)

 
 PDF Download Permissions and Reprints All articles of this category


Abstract

The molecular structure of the cytotoxic cembranolide uprolide G acetate (UGA) was proposed in 1995 and subsequently revised in 2000 on the basis that NMR data for UGA were very similar to those of a synthetic analogue that was unambiguously confirmed by X-ray diffraction analysis. Our synthetic studies of UGA suggested that the revised structure for UGA was still incorrect. Therefore, two new possible structures for UGA were proposed based on comprehensive NMR data analysis. The proposed structures were synthesized in 33 steps by exploitation of Achmatowicz rearrangement, ring-closing metathesis, and Sharpless asymmetric dihydroxylation as the key steps. Their analysis led to the identification of the correct structure for UGA. The success of structural revision of UGA illustrated well the importance of the interplay between NMR data analysis and chemical synthesis.


#

Key words

structure elucidation - cembranolide - uprolide - total synthesis - natural products

Natural products have played a key role in drug discovery[1] and in advancing organic synthesis.[2] The determination of their molecular structures including relative and absolute configurations is of paramount importance because they provide the basis for their subsequent studies. The structures of natural products are usually proposed on the basis of extensive analysis of their spectroscopic data, particularly of NMR spectra, with modern structural elucidation strategies/methods.[3] However, there remain many natural products with structures that have been misrepresented because of their high structural complexity and/or unprecedented structural motifs. In this regard, chemical synthesis has been widely recognized as an essential tool for the identification of structural misassignments and for providing proof of structural revisions.[4] In some cases, several cycles of chemical synthesis and structural revisions are required to reach the correct structures for the corresponding natural products. Therefore, it is very important that reasonable structures are proposed based on meticulous analysis of spectroscopic data available from natural and synthetic compounds and that a flexible and robust synthetic route for chemical synthesis is developed. Herein, we highlight a recent successful example: structural revision of uprolide G acetate,[5] to illustrate the importance of the effective interplay of NMR data analysis and chemical synthesis.

Zoom Image
Liangyu Zhu(left) obtained his B.Sc. degree in 2011 from Wuhan University, Wuhan, P. R. of China and subsequently joined Prof. Tong’s research group as a Ph.D. candidate at the Hong Kong University of Science and Technology. His current research is concentrated on the ­total synthesis of cembranolide uprolides.
Rongbiao Tong (right) received his B.Sc. in 2000 and M.Sc. in 2003 from Hunan University, Changsha, P. R. of China and then joined the ­research group of Prof. Frank E. McDonald at Emory University, USA, where he was awarded a Ph.D. degree in 2008. After working with Prof. Amos B. Smith, III at the University of Pennsylvania as a postdoctoral fellow (2008–2011), he started his independent research career in 2011 as assistant professor at the Hong Kong University of Science and Technology (HKUST). His current research interest focuses on total synthesis of complex biologically active natural products by exploitation of oxidative dearomatization of phenols and furfuryl alcohols.

Uprolides D–G [Scheme [1] (a)] were isolated by Rodriguez and co-workers in 1995 from gorgonians of Eunicea mammosa as structurally novel cytotoxic α-methylene-γ-lactone- (αmγl-) bearing cembranolides that were characterized by the presence of a 4,7-oxa-bridged cyclic ether (e.g., tetrahydrofuran, THF), which was unprecedented at that time.[6] The promising biological activities and novel structures of these natural products have surprisingly not received much attention from synthetic communities,[7] even though the corresponding cembranolides with simpler structure and less functionality such as crassin have attracted considerable interest in their chemical synthesis and bioactivity studies.[8]

Zoom Image
Scheme 1 (a) Uprolides D–G; (b) Structural revisions of uprolide G acetate and uprolide F diacetate by Rodriguez in 2000

In 2000, Rodriguez et al. undertook the synthesis of uprolides D–G analogues from the natural product eupalmerin acetate, which led them to make structural revisions of uprolide G acetate (UGA) and uprolide F diacetate (UFD) as the corresponding tetrahydropyrans 1a and 1b, respectively [Scheme [1] (b)].[9] Specifically, treatment of eupalmerin acetate (4) with p-TSA–H2O in THF at reflux temperature for 5 h gave diol 5 as the major product (55%), which was subjected to epoxidation with m-CPBA (49%) and acid-promoted regio- and stereoselective epoxide opening to provide compounds 2 (38%, confirmed by X-ray analysis) and 3 (50%). To their surprise, the 1H and 13C NMR spectra data of synthetic tetrahydropyran 2 were remarkably similar to those of UGA and UFD, whereas the spectra data of the tetrahydrofuran 3 clearly differ from those of UGA and UFD (Figure [1]). The only significant variations of 2 and UGA were found at C4, C5, and C18 in 13C NMR and H-C5, H-C6, and H-C9 in 1H NMR, which were believed to arise from the opposite stereochemistry at C4. On the basis of these synthetic studies, reinterpretation of spectral data of UGA and UFD, and careful comparison of 1H and 13C NMR data of 2 and 3 with those of natural UGA and UFD, Rodriguez et al. concluded that the structures reported for UGA and UFD should be revised as tetrahydropyrans 1a and 1b, respectively. Apparently, the chemical synthesis (semisynthesis in this example) played an indispensable role in identification of the constitutional structure misassignments of UGA and UFD, which otherwise would not be detected. On the other hand, the proposed, plausible structures (1a and 1b) could not be prepared or confirmed by this semi-synthesis approach because the C4 stereochemistry derived from eupalmerin acetate (4) could not be readily inverted without inversion of the C3 chiral center on this substrate or intermediate 4. In addition, it is questionable that the NMR differences between compound 2 and UGA/UFD arise solely from inversion of C4 stereochemistry and that the 8-O-methyl group of 1a or 8-O-acetyl group of 1b would not cause significant deviations of chemical shifts from 2. The inherent limitation of semi-synthesis and the uncertainty of true structures require a total synthesis with flexibility to access both stereochemistry of C4.

Zoom Image
Figure 1 Chemical shift variations between uprolide G acetate (UGA) and synthetic analogues 2 and 3

To verify the revised structures (1a and 1b) for the natural UGA and UFD with the intention of developing a flexible and robust synthetic route for these underexplored cytotoxic αmγ-cembranolides, we in 2012 started a synthetic program directed to the total synthesis of UGA (1a) as our initial goal. After attempting several unsuccessful macrocyclization methods, we finally found that ring-closing metathesis (RCM) could be employed to reliably and reproducibly forge the 14-membered cembrane skeleton in excellent yield (Scheme [2]), and some key transformations are highlighted herein.

Zoom Image
Scheme 2 Total synthesis of the proposed structure (1a) of uprolide G acetate

Our synthesis began with preparation of enantiomerically pure furfuryl alcohol 7 from furfuryl aldehyde 6 using Noyori asymmetric transfer hydrogenation (Noyori ATH) to install the absolute stereochemistry at C3.[10] Achmatowicz rearrangement[11] of 7 followed by Kishi reduction[12] and Pd-catalyzed hydrogenation provided tetrahydropyran core 8 of UGA with remarkable efficiency. Nucleophilic addition of methyl nucleophile to tetrahydropyranone 8 was found to be poorly diastereoselective,[13] leading to a 1:1.1 mixture of diastereomers. However, this, in principle, permitted us to access both stereoisomers at C4 for NMR spectra studies because the C4 stereochemistry was believed to cause the chemical shift deviations of UGA from 2. It should be noted that the determination of the absolute configuration of the tertiary alcohol (9 and 11) was not trivial because the NMR techniques using Mosher’s ester could not provide reliable and conclusive results due to possible dehydration and/or epimerization.[14] The desired alcohol 9 was carried forward with five steps to provide ketone 10, to which CeCl3-mediated nucleophilic addition of vinyl Grignard delivered a 7:1 mixture of diastereomers favoring Felkin–Anh controlled desired product 11. The ten-step elaboration of 11 provided conjugate aldehyde 12 for the subsequent Abiko–Masamune anti-aldol,[15] which installed the stereochemistry of both the methyl group at C12 and the secondary alcohol at C13. The latter stereochemical information was transferred to C1 with C1–C15 bond formation through Johnson–Claisen rearrangement.[16] Sharpless asymmetric dihydroxylation[17] of 14 and concomitant lactonization yielded the γ-lactone 15, which was elaborated in ten steps to the proposed structure (1a) of UGA, featuring a key, highly efficient ring-closing metathesis.[18] The synthetic route built on tremendous efforts and attempts represented a new synthetic strategy for the 14-membered αmγ-cembranolides.

To our disappointment, the NMR spectra data of our synthetic 1a were not in agreement with those of the natural UGA, as shown in Figure [2]. This finding clearly suggested that the purported structure (1a) for UGA was incorrect. To approach the true structure of UGA, we re-examined the reported NMR data of UGA and the synthetic analogue 2 and made a comprehensive analysis of these spectra data with respect to the structures of 2 and 1a. The most significant differences occurred around C4 (i.e., C3, C4, C5, and C18) for analogue 2 and surrounding both C4 and C8 (i.e., C7, C9, and C19) for 1a, which implied that UGA should have an O-methyl group at C4 instead of C8 and a free alcohol at C8, not C4. This rationalization was reinforced when considering that the substituent (O-Me or O-Ac) would usually change the local chemical environment (e.g., electron density) and consequently change the chemical shifts. In addition, the substituent would give rise to different conformations for macrocycles (in particular, the conformationally mobile 14-membered cembranolides[19]), leading to variations of chemical shifts. This deductive analysis led us to propose two possible structures for UGA: 18a and 18b (Scheme [3]).

Zoom Image
Figure 2 Chemical shift variations between uprolide G acetate and the synthetic 1a and 2

Before setting out to undertake the painstaking 42-step total synthesis, we performed further studies of the substituent effect[20] of the methyl group on 13C NMR chemical shifts by using the three sets of spectra data reported for UGA, 2, and 1a. When comparing the 13C NMR data of 2 and 1a, it was evident that the O-methyl substituent at C8 caused a positive effect on C8 (+4.2 ppm, positive β-effect) and a negative effect on C7 (–5.1 ppm, negative γ-effect) and C9 (–6.1 ppm, negative γ-effect) (Scheme [3]). Therefore, if 18a was UGA, similar substituent effects should be expected on C4: positive β-effect and negative γ-effect. Apparently, both of these effects were consistently observed in terms of sign and magnitude (Scheme [3]). On the other hand, 18b was less likely to be the natural UGA because abnormal γ-effect at C3 and C5 was observed: +1.0 ppm for C3 and +1.2 ppm for C5. In addition, the C4 stereochemistry should be identical to that of 2 if taking considerations of possible biogenesis and the biomimetic synthetic studies conducted by Rodriguez, as shown in Scheme [1] (b).

Zoom Image
Scheme 3Possible structures for uprolide G acetate. S effect: Substituent effect

The combination of these findings suggested that 18a should be the correct structure for UGA. However, the reported observation of an NOE effect for H(C3)–H(Me18) cast doubt on the exclusion of 18b as a possible structure for UGA. In this case, only total synthesis of both 18a and 18b could resolve the uncertainty and therefore reach the true structure for the natural UGA.

The confidence on 18a being the true structure of UGA based on this comprehensive NMR data analysis and the robust and flexible synthetic route developed previously for 1a led us to start its total synthesis (Scheme [4]) from 9a by using a similar synthetic route to that developed for 1a, whereas its diastereomer 18b could also be prepared from the corresponding diastereomeric 9b in the less likely event of 18b being the natural UGA. O-Methylation of 9a and 9b with NaH/MeI and a five-step sequence to stereoselectively install the tertiary alcohol at C8 with subsequent protection as triethylsilyl ethers, provided 19a and 19b, respectively. The elaboration of 19a and 19b to 18a and 18b, respectively, proceeded smoothly over 24 steps, demonstrating the robustness of our previous synthetic route. Single-crystal X-ray diffraction analysis of 18b implied that there were no unexpected transformations or epimerization in the course of synthesis. Most excitingly, we found that the NMR spectra data of our synthetic 18a were identical to those of the natural UGA, which substantiated our constitutional structure revision and constituted the first asymmetric total synthesis of UGA. The successful confirmation of our structural revision for UGA by total synthesis could be attributed to an effective interplay between the critical and comprehensive NMR data analysis and a flexible chemical synthesis, which have been implemented three times in this special case. The semisynthesis and NMR data analysis by Rodriguez led to the first structural revisions from THF-containing uprolides to the corresponding THP-embedded uprolides and our first total synthesis (Scheme [2]), and subsequent meticulous NMR data analysis resulted in a second round of structural revisions; the structure was eventually confirmed by our second chemical synthesis. In this context, chemical synthesis is clearly proved to be an effective way to detect structural misassignments and to confirm structural revision of natural products, whereas effective NMR data analysis can reduce the structural space in which to search (e.g., with respect to isomers) considerably when natural products are structurally misrepresented.

Zoom Image
Scheme 4Synthesis of uprolide G acetate and 4-epi-uprolide G acetate

In summary, we have described the structural revision of uprolide G acetate and highlighted the importance of effective interplay of NMR data analysis and chemical synthesis for structural revisions. The problem-solving strategy reported in this paper might be applicable to many other cases for which the wrong purported structures of natural products were targeted by total synthesis. Therefore, some specific suggestions/procedures are given here: i) diagrams (cf. Figure [1] and Figure [2]) of NMR chemical shift differences between the natural product and the synthetic (analogue or congener) compounds should be made; ii) the major NMR differences with respect to stereochemistry and substituent effects (β- and γ-effects) on chemical shifts should be rationalized; iii) the biogenesis of the compound should be considered when making stereochemistry assignments; and iv) a flexible and robust synthetic route should be developed.


#

Acknowledgment

This research was financially supported by HKUST, the Research Grant Council of Hong Kong (GRF 606113 and GRF 16305314), and by the National Natural Science Foundation of China (NSFC 21472160).



   Permissions and Reprints All articles of this category
Zoom Image
Liangyu Zhu(left) obtained his B.Sc. degree in 2011 from Wuhan University, Wuhan, P. R. of China and subsequently joined Prof. Tong’s research group as a Ph.D. candidate at the Hong Kong University of Science and Technology. His current research is concentrated on the ­total synthesis of cembranolide uprolides.
Rongbiao Tong (right) received his B.Sc. in 2000 and M.Sc. in 2003 from Hunan University, Changsha, P. R. of China and then joined the ­research group of Prof. Frank E. McDonald at Emory University, USA, where he was awarded a Ph.D. degree in 2008. After working with Prof. Amos B. Smith, III at the University of Pennsylvania as a postdoctoral fellow (2008–2011), he started his independent research career in 2011 as assistant professor at the Hong Kong University of Science and Technology (HKUST). His current research interest focuses on total synthesis of complex biologically active natural products by exploitation of oxidative dearomatization of phenols and furfuryl alcohols.
Zoom Image
Scheme 1 (a) Uprolides D–G; (b) Structural revisions of uprolide G acetate and uprolide F diacetate by Rodriguez in 2000
Zoom Image
Figure 1 Chemical shift variations between uprolide G acetate (UGA) and synthetic analogues 2 and 3
Zoom Image
Scheme 2 Total synthesis of the proposed structure (1a) of uprolide G acetate
Zoom Image
Figure 2 Chemical shift variations between uprolide G acetate and the synthetic 1a and 2
Zoom Image
Scheme 3Possible structures for uprolide G acetate. S effect: Substituent effect
Zoom Image
Scheme 4Synthesis of uprolide G acetate and 4-epi-uprolide G acetate