Recent applications of ring-rearrangement metathesis in organic synthesis

  1. Sambasivarao Kotha1,
  2. Milind Meshram1,
  3. Priti Khedkar2,
  4. Shaibal Banerjee3 and
  5. Deepak Deodhar1

1Department of Chemistry, Indian Institute of Technology Bombay, Powai, Mumbai-400 076, India. Fax: (+91)-22-2572-3480 2576 7152; Phone: (+91)-22-2576-7160
2Department of Chemistry, Guru Nanak Khalsa College of Arts, Science & Commerce, Matunga, Mumbai–400 019, India
3Department of Applied Chemistry, Defence Institute of Advanced Technology (DU), Girinagar, Pune–411025, Pune, India

  1. Corresponding author email

This article is part of the Thematic Series "Progress in metathesis chemistry II".

Guest Editor: K. Grela
Beilstein J. Org. Chem. 2015, 11, 1833–1864. doi:10.3762/bjoc.11.199
Received 13 Jun 2015, Accepted 17 Sep 2015, Published 07 Oct 2015

Abstract

Ring-rearrangement metathesis (RRM) involves multiple metathesis processes such as ring-opening metathesis (ROM)/ring-closing metathesis (RCM) in a one-pot operation to generate complex targets. RRM delivers complex frameworks that are difficult to assemble by conventional methods. The noteworthy point about this type of protocol is multi-bond formation and it is an atom economic process. In this review, we have covered literature that appeared during the last seven years (2008–2014).

Keywords: Diels–Alder chemistry; green chemistry; natural products; olefin metathesis; polycycles; ring-rearrangement metathesis

Introduction

Transition metal–carbene complexes (Figure 1) introduced during the last two decades have changed the landscape of organic synthesis. Armed with these advances, olefin metathesis has become a staple in retrosynthesis. Metathesis protocols such as ring-closing metathesis (RCM), cross-metathesis (CM), and enyne metathesis (EM) have gained popularity in the synthesis of complex molecules. Ring-rearrangement metathesis (RRM) involves a tandem process, where the ring-opening metathesis (ROM) and the RCM sequence occur in tandem to generate complex end products (Figure 2). Several demanding structures related to natural products and non-natural products were synthesized by RRM. However, a limited number of papers appeared dealing with RRM due to the complexity involved in designing the required precursors suitable for RRM. There are several factors which facilitates the RRM. Among them, the release of ring strain is the main driving force. For example, with bicyclo[2.2.1]heptene systems, RRM produce less strained end products. A general mechanism for the RRM process is shown in Figure 3 [1,2]. During RRM the stereochemical information is transformed from the substrate to the product. Interestingly, RRM is applicable to mono- and polycyclic systems of varying ring sizes. The outcome of the RRM process depends on the selection of the protecting groups, reaction conditions, and electronic properties of substrates involved. Oligomerization is a common side reaction in the RRM and external olefins such as ethylene prevents unwanted oligomerization processes. For earlier work related to the RRM readers may refer to excellent reviews available in the literature [3-6].

[1860-5397-11-199-1]

Figure 1: Ruthenium alkylidene catalysts used in RRM processes.

[1860-5397-11-199-2]

Figure 2: General representation of various RRM processes.

[1860-5397-11-199-3]

Figure 3: A general mechanism for RRM process.

Review

Cyclopropene systems

Cyclopropene derivatives are highly strained systems and they are ideal candidates for the RRM process. In this context, Zhu and Shi [7] have reported the ring-closing enyne metathesis (RCEM) of small-rings such as cyclopropenes by employing the Grubbs’ first-generation (G-I) catalyst. They have reported a new tandem ROM–RCM–CM sequence starting with 1,6-cyclopropenynes 16 with a wide variety of substituted olefins. To this end, the required building block 16 has been prepared with the aid of a carbene insertion reaction. Further, this cyclopropene system 16 was subjected to RRM in the presence of catalyst 1 to generate 3-pyrroline derivatives 18a,b using simple starting materials in a single step (Scheme 1).

[1860-5397-11-199-i1]

Scheme 1: RRM of cyclopropene systems.

A wide range of heterocycles have been assembled by RRM. When a substituted cyclopropene such as 19 (or 20) was treated with catalyst 2 in the presence of ethylene (24) the required heterocycle 22 (or 23) was obtained in moderate to good yield (Scheme 2) [8]. Allyl ethers 25a,b and 26a,b were reacted with catalyst 2 to deliver the corresponding dihydrofurans (27a,b and 28a,b) in excellent yields (82–92%). Involvement of acrylates 29a,b delivered lactones 30a,b in moderate yields (30a 41%, 30b 50%) upon treatment with catalyst 2 in dichloromethane at reflux conditions. However, 29a generated lactone 30a in 65% yield when the metathesis was performed using Grela’s catalyst 7. Pyrrolines were produced in excellent yields by RRM of sulfonamides 31a,b using the catalyst 2 under dichloromethane reflux conditions (32a 99%, 32b 70%) (Scheme 3). Five-membered heterocycles such as 34 and a seven membered heterocycle 35 in 40:60 ratio (97%) were formed by RRM of cyclopropenylcarbinyl ether 33 with catalyst 2 (Scheme 4).

[1860-5397-11-199-i2]

Scheme 2: RRM of cyclopropene with catalyst 2. (i) catalyst 2 (2.5 mol %), ethylene (24, 1 atm), (ii) toluene (c = 0.02 M), reflux, (iii) CH2Cl2 (c = 0.02 M), reflux, (iv) C6H6 (c = 0.01 M), reflux. (a) without ethylene (24); (b) with ethylene (24).

[1860-5397-11-199-i3]

Scheme 3: RRM of various cyclopropene derivatives with catalyst 2. (i) catalyst 2 (2.5 mol %), CH2Cl2 (c = 0.1 M), reflux, (ii) (a) catalyst 2 (2.5 mol %), toluene (c = 0.1 M), reflux, (b) catalyst 7, toluene (c = 0.1 M), reflux.

[1860-5397-11-199-i4]

Scheme 4: RRM of substituted cyclopropene system with catalyst 2.

Cyclobutene systems

Cyclobutene is also highly strained and prone to RRM very easily. Maougal and co-workers synthesized 3,3’-bipiperidine and 3,3’-bis(1,2,3,6-tetrahydropyridine) systems through a RRM sequence [9]. In this context, they have identified compound 38 as the key starting synthone, easily prepared from 36 via an N-allylation sequence. Next, diallyl compound 38 was treated with catalyst 2 to deliver the expected bipiperidine derivative 39 in 60% yield. Further, this protocol has been extended to various oxygenated systems (Scheme 5).

[1860-5397-11-199-i5]

Scheme 5: RRM of cyclobutene system with catalyst 2.

Snapper and White [10] have reported a new and efficient method to various medium size bicyclic systems. Here, the RRM strategy has been employed with catalyst 2 starting with various cyclobutene systems containing an alkene tether (e.g., 40, 42, and 44) to generate bicyclic systems such as 41, 43, and 45 (Scheme 6).

[1860-5397-11-199-i6]

Scheme 6: RRM approach to various bicyclic compounds.

The erythrina alkaloids are known to exhibit sedative, hypotensive and neuromuscular activity. This alkaloid skeleton consists of a tetracyclic spiroamine framework and synthetic chemists consider it as a challenging target. Simpkins and co-workers [11] have used the RRM sequence tactically to assemble the erythrina skeleton. To this end, they have identified cyclobutene derivative 48 as a useful synthone for RRM. The cyclobutene derivative 46 has been extended via Grignard addition followed by cyclization reaction. Later, cyclobutene derivative 48 was treated with catalyst 1 in the presence of ethylene (24) under high dilution conditions to deliver the tetracyclic compound 49 in 62% yield (Scheme 7).

[1860-5397-11-199-i7]

Scheme 7: RRM approach to erythrina alkaloid framework.

To assemble 5-F2-isoprostanes, lipid oxidation metabolites, various functionalized cyclobutene derivatives were subjected to a RRM sequence [12]. Cyclobutene derivative 50 in the presence of the catalyst 1 delivered lactone 51 as a mixture of isomers (3:1) in 37% yield. When the substrate was modified as in 52, the RCM product was not formed; however, compound 52 gave the ring-opened product with ethylene (24) in low yield. Further, the ROM homodimer was obtained in 17% yield in the presence of ethylene (24) with the aid of catalyst 4 (Scheme 8).

[1860-5397-11-199-i8]

Scheme 8: ROM–RCM sequence to lactone derivatives.

Pattenden and co-workers [13] have described a novel synthesis of (+)-Z-deoxypukalide using substituted butenolide intermediate 58. Interestingly, it was synthesized starting with cyclobutene ester 55 involving ROM–RCM and CM protocols. In this regard, the cyclobutene ester was subjected to a ROM–RCM and CM protocol under conditions with catalyst 2 in the presence of 2-methylpropenol 57 to afford the required butenolide intermediate 58 in 57% yield (Scheme 9).

[1860-5397-11-199-i9]

Scheme 9: RRM protocol towards the synthesis of lactone derivative 58.

An asymmetric synthesis of humulanolides is achieved by a RRM approach. In this context, Li and co-workers [14] prepared the key precursor 60 in five steps from commercially available starting material 59. Later, the cyclobutene derivative 60 was treated with catalyst 5 under toluene reflux conditions to give the expected RRM cascade product, i.e., asteriscunolide D (61) in 36% yield along with the dimer 62 (7%). Interestingly, they also found asteriscunolide D as a useful synthone for the synthesis of asteriscunolides A−C (Scheme 10).

[1860-5397-11-199-i10]

Scheme 10: RRM protocol towards the asymmetric synthesis of asteriscunolide D (61).

In several instances RRM has proved to be a useful strategy for the construction of 12- to 16-membered macrolides [15]. In this regard, ester 65 was prepared from the corresponding allylic alcohol 63 by esterification with the anhydride 64 derived from cyclobutene. Later, the ester 65, on treatment with the catalyst 1 under toluene reflux conditions followed by treatment with the catalyst 2 furnished the macrolide-butenolides 66 in 42–48% yields via RRM with E-selectivity at the macrocyclic double bond. Along similar lines, compound 65f was treated with catalyst 1 in refluxing toluene followed by treatment with catalyst 2 to deliver desmethylmanshurolide 67 in 44% yield (Scheme 11).

[1860-5397-11-199-i11]

Scheme 11: RRM strategy towards the synthesis of various macrolide rings.

Cyclopentene systems

In RRM with cyclopentene systems, the release of ring strain is a less important contributor to the driving force of the reaction. However, unfavorable interaction of vicinal or proximal substituents may be minimized in the rearranged product. In this context, Blechert and co-workers [16] demonstrated the first enantioselective total syntheses of virgidivarine and virgiboidine by employing an intramolecular ene–ene–yne domino RRM protocol in combination with an oxidative C–C bond cleavage. This protocol opens-up new opportunities for the construction of intricate dipiperidine-based targets in a stereoselective manner (Scheme 12).

[1860-5397-11-199-i12]

Scheme 12: RRM protocol to dipiperidine system.

Lee and Li [17] disclosed a highly distereoselective RRM approach starting with cyclopentene derivatives. In this regard, the cyclopentene derivative 72 was treated with the catalyst 2 in the presence of ethylene (24) to generate the required cyclohexene-based product 73. The total synthesis of spiropiperidine alkaloid nitramine was proved to be efficient by this methodology (Scheme 13).

[1860-5397-11-199-i13]

Scheme 13: RRM of cyclopentene system to generate the cyclohexene systems.

In 2004, Ni and Ma [18] have described the synthesis of bicyclic compounds 75 and 76 by adopting a metathesis protocol with catalysts 1 and 2 in good yields, but the product ratio is catalyst-dependent. In this context, when the cyclopentene derivative 74 gave 75 and 76 (1:5, 75%) with catalyst 1; whereas, the catalyst 2 produced 75 and 76 in 85% yield (12:1). Here, they have shown the thermodynamically favored RRM leads to the formation of 75, while kinetically favored RCM gave the product 76 (Scheme 14).

[1860-5397-11-199-i14]

Scheme 14: RRM of cyclopentene system 74.

Cyclohexene systems

Banti and co-workers have reported a tandem metathesis sequence with the aid of catalysts 1 and 2 starting with cyclohexene and norbornene systems containing allylamino moieties [19]. When the reaction was carried out in the presence of catalyst 2; RRM product 79 was observed in 29% yield along with the RCM product 78 in 71% (Scheme 15).

[1860-5397-11-199-i15]

Scheme 15: RRM approach to compound 79.

Burnell and co-workers [20] have demonstrated the RRM of unsaturated spirocycles with two alkenyl chains by employing catalyst 2 to generate a unsaturated spiro-fused tricyclic system. In this context, the compounds 80 and 81 were subjected to RRM with catalyst 2 to furnish exclusively fused tricyclic systems 83a and 83b in 85% and 61% yields, respectively. Substitution on the cyclohexene system as in compound 82 did not deliver the RRM product (Scheme 16).

[1860-5397-11-199-i16]

Scheme 16: RRM approach to spirocycles.

Pyran systems

Donnard and co-workers [21] have accomplished a RRM approach for assembling complex heterocycles by employing simple starting materials. They have studied the RRM of dihydropyrans and dihydrofurans and this approach was found to be useful for the synthesis of non-classical saccharides. The synthesis of unusual di- or trisaccharides and related systems are also accessible by this approach. The required building block 85 has been prepared from compound 84 by allylation with allyl bromide (37). Later, the pyran derivative 85 was treated with catalyst 2 to generate the bicyclic system 86 (73%) (Scheme 17).

[1860-5397-11-199-i17]

Scheme 17: RRM approach to bicyclic dihydropyrans.

They also demonstrated a RCM–ROM–RCM cascade using a strain-free allyl heterocycle as useful starting material [22]. The required building blocks such as 90a–c were prepared from compound 87. Later, treatment of 90b with catalyst 2 gave the expected RRM product 91b (83%), whereas compound 90a gave the rearranged product 91a in 8% yield. On the other hand, when compound 90c was reacted with catalyst 2 the rearranged product was not formed. Here, they have demonstrated that the success of the reaction depends on electronic and stereochemical factors. Moreover, the synthesis of unusual polydeoxydisaccharides could be achieved starting with these simple starting materials. Similarly, 93 has been obtained by the RRM of compound 92 (Scheme 18).

[1860-5397-11-199-i18]

Scheme 18: RCM–ROM–RCM cascade using non strained alkenyl heterocycles.

Eustache and co-workers [23] have reported a novel ROM–RCM–ROM–RCM cascade involving a simple heterocycle as a useful precursor for the RRM protocol. To this end, the required precursor 96 was synthesized from 94 in two steps. Next, 96 was treated with catalyst 2 to generate the expected RRM product 97 in 68% yield. Further, this approach is useful for the preparation of polyunsaturated trisaccharides (Scheme 19).

[1860-5397-11-199-i19]

Scheme 19: First ROM–RCM–ROM–RCM cascade for the synthesis of trisaccharide 97.

Mori and co-workers [24] have used the RRM protocol starting with enyne 98 using catalyst 2 in the presence of ethylene (24) to generate the dimerized 16-membered ring product 101 in 57% yield, which was generated by a RRM–dimerization sequence and its monomer 100 in 14% yield along with 99 in 26% yield (Scheme 20).

[1860-5397-11-199-i20]

Scheme 20: RRM of cyclohexene system.

Bicyclo[2.2.1]heptene derivatives

Holtsclaw and Koreeda [25] have explored a chemoselective RRM of the enone containing norbornene system such as 102. To this end, the spironorbornene 102 was subjected to a RRM sequence under the influence of catalyst 1 to deliver the RRM product 103 and the RCM product 104 in a 99:1 ratio. When norbornene derivative 102 was treated with catalyst 2 tricyclic compound 103 and spironorbornene derivative 104 were obtained in a 22:78 ratio (Scheme 21).

[1860-5397-11-199-i21]

Scheme 21: RRM approach to tricyclic spirosystem.

Aubé and co-workers [26] have accomplished the asymmetric synthesis of the dendrobatid alkaloid 251F by employing a RRM as the key step. The required building block 108a has been synthesized from enone 107 via a RRM protocol. When enone 107 was exposed to catalyst 1 in the presence of ethylene (24) the RRM product 108a was obtained in 93% yield. Further, this bicyclic building block 108a has been successfully utilized in the synthesis of the dendrobatid alkaloid 251F (Scheme 22).

[1860-5397-11-199-i22]

Scheme 22: RRM approach to bicyclic building block 108a.

Phillips and Henderson [27] have demonstrated the synthesis of aburatubolactam A (113) by using a tandem ROM–RCM sequence as the key step. To this end, the required key building block, the bicyclo[3.3.0]octene ring system 108b has been synthesized by a RRM sequence via catalyst 1 starting with the functionalized bicyclo[2.2.1]heptene system 107. Thus, the Diels–Alder (DA) reaction of ketone 109 with cyclopentadiene (111) in the presence of MacMillans catalyst 110 gave bicyclic ketone 112 in 65% yield. Then, ketone 112 was transformed into enone 107 in 80% yield by adopting the known oxidation protocol. Later, enone 107 was treated with catalyst 1 under ethylene (24) atmosphere to deliver the required bicyclo[3.3.0]octane derivative 108b in 90% yield (Scheme 23).

[1860-5397-11-199-i23]

Scheme 23: ROM–RCM protocol for the synthesis of the bicyclo[3.3.0]octene system.

Shibatomi and co-workers have reported an enantioselective DA reaction of β-fluoromethylacrylate under the influence of the chiral Lewis acid-activated catalyst, oxazaborolidine to generate the difluoromethylated cycloaddition endo-product 114 (99% ee). Further, it was used to prepare the required enone 115. Later, enone 115 was subjected to a RRM protocol by employing catalyst 1 in the presence of ethylene (24) to generate the bicyclic enone 116 in 53% yield (Scheme 24) [28].

[1860-5397-11-199-i24]

Scheme 24: RRM protocol to bicyclic enone.

In 2005, Funel and Prunet have disclosed the synthesis of fused tricyclic systems by employing a RRM protocol [29]. For example, the bicyclic system 117 was treated with catalyst 2 to generate the rearranged tricyclic system 118. This strategy has been extended with higher analogues (Scheme 25).

[1860-5397-11-199-i25]

Scheme 25: RRM protocol toward the synthesis of the tricyclic system 118.

Kotha and Ravikumar [30] have utilized the RRM and the enyne RRM to generate various polycyclic scaffolds. In this context, enones, such as 121a and 121b were assembled easily from dicyclopentadiene derivative 119. Later, these componds were subjected to a RRM to generate the tricyclic enones 122a and 122b, respectively. To this end, compound 121a was treated with catalyst 2 under ethylene (24) atmosphere to deliver the tricyclic enone 122a in 75% yield. Similarly, the tricyclic compound 122b (50%) was obtained under the influence of catalyst 5 in the presence of ethylene (24, Scheme 26).

[1860-5397-11-199-i26]

Scheme 26: RRM approach toward the synthesis of the tricyclic enones 122a and 122b.

Along similar lines, the oxa analog 125 was obtained by RRM of 124 using catalyst 1 under ethylene (24) atmosphere. Interestingly, the diene building block 128, produced by employing an enyne-ring rearrangement metathesis (ERRM) sequence, was subjected to a DA reaction in refluxing toluene with various dienophiles such as dimethyl acetylenedicarboxylate (DMAD, 129) to generate the tetracyclic system 130 (Scheme 27).

[1860-5397-11-199-i27]

Scheme 27: Synthesis of tricyclic and tetracyclic systems via RRM protocol.

To design synthetically challenging oxa-bowls, Kotha and Ravikumar [31] have utilized the RRM and ERRM of extended norbornene systems. The key building blocks such as 133 and 134 were prepared from a readily available DA adduct 131 derived from cyclopentadiene (111) and 1,4-benzoquinone. The diol 132 was produced by reduction of 131 in an efficient manner. To this end, the bis-O-allylated compound 133 was prepared by an allylation sequence using allyl bromide (37) in the presence of NaH starting with the diol 132, whereas compound 134 was derived via bis-O-propargylation of compound 132 using propargyl bromide (126) under similar reaction conditions. Later, these compounds (133 and 134) were subjected to RRM and ERRM protocols under the influence of catalyst 1 in the presence of ethylene (24) to generate the tetracyclic systems 135 (100%) and 136 (76%), respectively. Moreover, this strategy can easily be extended to other complex systems (Scheme 28).

[1860-5397-11-199-i28]

Scheme 28: RRM protocol towards the synthesis of tetracyclic systems.

Banti and co-workers have described a RRM with catalysts 1 and 2 by using an aminopropargylated norbornene system as a starting material [19]. In this reaction, three possible products were observed by employing either catalyst 1 or 2. The norbornene derivative 137 was subjected to a RRM protocol under the influence of catalyst 1 in the presence of ethylene (24) to obtain the expected product 138 in 41% yield (Table 1, entry 1) along with the cis- and trans-monocyclized products 139 and 140. Further, NMR spectroscopic studies showed the presence of products 139 and 140 as a mixture of isomers, and it was difficult to purify this mixture by conventional separation techniques (Scheme 29).

Table 1: RRM of propargylamino derivative.

Entry Cat (mol %) Solvent T (°C) time (h) Conv. 138 (yield %)
1 1 (5) CH2Cl2 25 6 100 41
2 1 (5) CH2Cl2 25 16 100 43
3 1 (5) + 2 (5) CH2Cl2 25 24 100 37
4 2 (5) toluene 60 24 0 0
[1860-5397-11-199-i29]

Scheme 29: RRM of the propargylamino[2.2.1] system.

Recently, Kotha and Gunta have reported a RRM to generate various polycyclic compounds using catalysts 1 and 2 [32]. The tetraallyl derivative, prepared from 142 by an allylation protocol, was subjected to a RRM sequence in the presence of the catalyst 1 to produce propellane derivative 144 containing an oxa-bowl moiety. In another sequence [33], the alkenylation of sulfone 145 gave the dialkenylated product 147 in 21% yield along with the monoalkenylated product. Later, the dialkenylated compound 147 was treated with catalyst 2 to give the tetracyclic compound 148 in 97% yield (Scheme 30).

[1860-5397-11-199-i30]

Scheme 30: RRM of highly decorated bicyclo[2.2.1] systems.

Along similar lines, Kotha and co-workers [34] prepared N-allylated compounds and subjected them to a RRM to produce the tricyclic aza compound 152 in an excellent yield. The required synthone 151 was prepared by employing a Beckman rearrangement followed by a N-allylation sequence. Later, it was reacted with catalyst 2 in the presence of ethylene (24) to deliver the expected tricyclic product 152 in 90% yield (Scheme 31).

[1860-5397-11-199-i31]

Scheme 31: RRM protocol towards fused tricyclic compounds.

Ghosh and Maity [35,36] reported a stereoselective route to functionalized tricyclic system present in umbellactal (153) via a RRM protocol starting with intricate norbornene derivatives. The tricyclic anhydride 154 was reduced to lactone 155 using sodium borohydride and then it was monoallylated to deliver an inseparable mixture of products 156 and 157 in 85% combined yield. Later, the mixture (156 and 157) was subjected to a RRM protocol under the influence of the catalyst 1 in the presence of ethylene (24) to yield a mixture (4:1) of tricyclic lactones 158 and 159 (70% yield). Next, the major product 158 was converted into 159 by isomerization via DBU in 82% yield. The cis-lactone 159 was found to be a core structural unit present in umbellactal (Scheme 32).

[1860-5397-11-199-i32]

Scheme 32: RRM protocol to functionalized tricyclic systems.

Ghosh and co-workers also reported [37] a short and efficient approach to a highly functionalized lactarane skeleton using RRM with appropriate norbornene systems. The strategy starts with the aldol condensation of aldehyde 160 with ester 161 in the presence of LDA to generate the required building block 162 in 78% yield. Later, the norbornene derivative 162 was subjected to a RRM sequence under the influence of the catalyst 1 in the presence of ethylene (24) to produce the rearranged product 163 in 65% yield (Scheme 33).

[1860-5397-11-199-i33]

Scheme 33: RRM approach to functionalized polycyclic systems.

Ghosh and co-workers have described an efficient route for the synthesis of the fused tricyclic system found in caribenol A by employing a RRM approach [38]. The steps employed here involve: a sequential aldol condensation of dihydrocarvone with norbornene 2-carboxaldehyde followed by a ROM–RCM of the resulting aldol product. The norbornene derivative 164 was subjected to a RRM using the catalyst 1 to produce the ROM product 165 exclusively. The ring-closure of the resulting ROM product 165 under the influence of the catalyst 2 led to the formation of the dimeric product. Alternatively, RCM of 165 under the influence of catalyst 5 generated the required tricyclic compound 166 in 45% yield (epimeric mixture at C-5). Interestingly, this tricyclic system was found as a core structural unit present in caribenol A (Scheme 34).

[1860-5397-11-199-i34]

Scheme 34: Sequential RRM approach to functionalized tricyclic ring system 166.

In another instance, they [39] achieved an efficient synthesis of the functionalized tricyclic ring system 171 in the context of the synthesis of the nonterpenoids schintrilactones A and B by a RRM approach of alkenylated norbornene derivative 170. They also reported an impressive set of example with complex norbornene systems. The required synthone 170, suitable for RRM, has been prepared from 167 in three steps. Later, compound 170 was treated with catalyst 2 in the presence of ethylene (24) to generate the desired tricyclic ring system 171 (94%), which is found to be a core structure of schintrilactones A and B (Scheme 35).

[1860-5397-11-199-i35]

Scheme 35: RRM protocol to functionalized CDE tricyclic ring system of schintrilactones A and B.

In 2012, Ghosh’s group [40] demonstrated a RRM approach towards the synthesis of a 7/5 fused system by using a bicyclo[2.2.1]heptene derivative via a sequential RRM approach. Moreover, they have studied the feasibility of a RRM protocol starting with highly substituted bicyclo[2.2.1]heptene and bicyclo[2.2.2]octene systems. Here, the silyl ether 172 was treated with catalyst 1 to give the ring-opened product 174. Next, the triene 174 was subjected to a RCM protocol in the presence of catalyst 2 to furnish the tricyclic product 176. Along similar lines, methyl substituted norbornene derivative 173 was treated with catalyst 1 in the presence of ethylene (24) to generate the ROM product 175, which was further subjected to a RCM using catalyst 2 to deliver the expected tricyclic system 177 (7/5 fused system) (Scheme 36).

[1860-5397-11-199-i36]

Scheme 36: Sequential RRM approach to 7/5 fused bicyclic systems.

A synthesis of fused medium-sized rings has been reported by Ghosh and co-workers [41] via a sequential diastereoselective DA reaction and a RRM protocol. A variety of sugar-based norbornene derivatives provide an entry to various functionalized bicyclic sugar derivatives containing 7–9 membered rings. To this end, compounds 178 and 181 were subjected to a ROM sequence with catalyst 1 in the presence of ethylene (24) followed by treatment with catalyst 2 under the same reaction conditions to give the RRM products 180 and 183, respectively, derived from the ROM products. Here, the norbornene derivatives 178a,b,d and 181a,c,d furnished the RRM products 180a,b,d and 183a,c,d, respectively. As expected, when the compounds 178c and 181b were subjected to metathesis under the influence of the catalyst 1, the RRM products (180c and 183b) were obtained respectively (Scheme 37).

[1860-5397-11-199-i37]

Scheme 37: Sequential ROM-RCM protocol for the synthesis of bicyclic sugar derivatives.

Along similar lines, compound 185 was reacted with cyclopentadiene in a DA fashion to deliver an inseparable mixture of adducts 186 and 187. Later, the ROM–RCM of this mixture of norbornene derivatives, gave the cis-syn-cis and cis-anti-cis 5-7-6 tricyclic systems 188 (60%) and 189 (26%), respectively, via the RRM approach (Scheme 38).

[1860-5397-11-199-i38]

Scheme 38: ROM–RCM sequence of the norbornene derivatives 186 and 187.

Ghosh’s group [42] disclosed an elegant approach to a highly functionalized bridged tricyclic system by employing a RRM approach involving catalyst 1. The required synthone 192 has been prepared from the allyl substituted norbornene derivative 190 in three-steps. Later, the keto derivative 192 was subjected to a RRM sequence via catalyst 1 to generate the bridged tricyclic system 193 in 40% yield (Scheme 39).

[1860-5397-11-199-i39]

Scheme 39: RRM approach toward highly functionalized bridge tricyclic system.

A novel approach to highly functionalized tricyclic systems such as 197 and 198 has been reported via a RRM protocol. In this context, the endo-aldehyde 160 was identified as a starting material in the synthetic sequence and it was transformed into enone 194 by treatment of vinyl Grignard 106 followed by Jones oxidation. Later, enone 194 was subjected to a DA reaction in the presence of cyclopentadiene (111) to deliver an inseparable mixture of cycloadducts 195 (endo,endo) and 196 (exo,exo) in a 1:2 ratio. Then, treatment of the cycloadducts 195 and 196 separately with catalyst 1 in the presence of ethylene (24) furnished the tricyclic compounds 197 (23%) and 198 (45%), respectively (Scheme 40). Analogously, they have also achieved the synthesis of angularly annelated carbocycles by employing the RRM protocol starting with appropriate norbornene derivatives [43].

[1860-5397-11-199-i40]

Scheme 40: RRM approach toward highly functionalized tricyclic systems.

Recently, Kotha and Ravikumar [44] have found a new route to various polycyclic compounds by employing the DA reaction and the RRM protocol as key steps. To this end, the key building block 202 has been prepared from 199 via Grignard addition followed by O-allylation. The double DA adduct 199 has been derived from cyclopentadiene and 1,4-benzoquinone. Next, compound 202 was exposed to catalyst 2 in the presence of ethylene (24) to generate the expected hexacyclic system 203 (70%) containing 10 stereogenic centres (Scheme 41).

[1860-5397-11-199-i41]

Scheme 41: Synthesis of hexacyclic compound 203 by RRM approach.

Sakurai and co-workers have successfully established an enantioselective synthesis of the C3-symmetric chiral trimethylsumanene through a Pd-catalyzed cyclotrimerization and the RRM protocol as key steps [45]. Here compound 207 reacted with catalyst 1 in the presence of ethylene (24) to deliver a mixture of ring-opened products. A sequential treatment with catalyst 2 resulted in a ring-closing product to deliver the expected hexahydrotrimethylsumanene 208 in 24% yield. When the tris-norbornene derivative 207 was treated with catalyst 2 in the presence of (Z)-oct-4-ene the required RRM product 208 was formed in 26% yield. Later, the expected chiral buckybowl 209 was assembled via aromatization of 208 in the presence of DDQ (Scheme 42).

[1860-5397-11-199-i42]

Scheme 42: RRM approach toward C3-symmetric chiral trimethylsumanene 209.

Design of intricate polyquinanes has been considered as a challenging task for synthetic chemists. To this end, Fallis and co-workers [46] have demonstrated an intramolecular Diels–Alder (IMDA) reaction followed by a RCM–ROM–CM cascade was found to be useful to assemble a linear triquinane framework. Microwave assisted IMDA reaction of cyclopentadiene derivative 210 performed in chlorobenzene at 201 °C under 310 psi pressure gave the required DA adduct 211. Later, the cycloadduct 211 was reacted with the catalyst 1 in the presence ethylene (24) to generate a linear cis-anti-cis triquinane derivative 212 (Scheme 43).

[1860-5397-11-199-i43]

Scheme 43: Triquinane synthesis via IMDA reaction and RRM protocol.

In search of new antibacterial drugs, Spring and co-workers [47] have designed a diversity-oriented approach to structurally diverse small molecules starting with solid-supported phosphonate 213. In this regard, they have shown the use of a RRM protocol to prepare the bicyclic product 218 as well as tricyclic product 217. To this end, the phosphonate ester 213 reacted with a wide variety of aldehydes 214 such as aryl, heteroaryl, and alkyl, etc. to produce α,β-unsaturated acylimidazolidinones 215. Next, the Evan’s asymmetric DA methodology involving a [4 + 2] cycloaddition of chiral bis(oxazoline) in the presence of Cu(OTf)2 was employed to furnish the required norbornene system 216. Later, it was converted into a lactam and then subjected to a RRM sequence with catalyst 2 in the presence of ethylene (24) to furnish the tricyclic product 217 as well as bicyclic product 218 (Scheme 44).

[1860-5397-11-199-i44]

Scheme 44: RRM approach to polycyclic compounds.

The bicyclo[3.3.0]octene system represent a core structural unit present in several natural products. Kimber and co-workers [48] have utilized a RRM approach to generate cis-fused bicyclo[3.3.0]octene derivatives. In this regard, various norbornenyl derivatives 219, 221, 223 and 225 were subjected to RRM by treatment with catalyst 2 in the presence of ethylene (24) to generate various bicyclo[3.3.0]octene derivatives such as 220, 222, 224, and 226 with high regioselectivity. The thermodynamic stability of the product is anticipated to play an important role in the observed regioselectivity of these transformations (Scheme 45).

[1860-5397-11-199-i45]

Scheme 45: RRM strategy toward cis-fused bicyclo[3.3.0]carbocycles.

In the course of the asymmetric synthesis of (−)-isoschizogamine, a bicyclic lactone 230 has been identified as a key building block. To this end, Fukuyama and co-workers [49] have used the RRM to generate the required building block 230. In this reaction, the required norbornene derivative 228 was prepared from epoxide 227 in two steps and later it was treated with the more reactive catalyst 6 in the presence of 1,6-heptadiene (229) to generate the required bicyclic lactone 230 (73%). In this process, 1,6-heptadiene (229) helps to enhance the rate of the reaction and to improve the yield. However, when the bicyclic system 228 was treated with catalyst 5 in refluxing benzene lactone 230 was obtained in 24% yield (Scheme 46).

[1860-5397-11-199-i46]

Scheme 46: RRM protocol towards the synthesis of bicyclic lactone 230.

Azanorbornene systems

7-Azanorbornene derivatives have been used to generate a wide variety of heterocyclic compounds via the RRM approach [50]. To this end, the azanorbornene derivative 231 was treated with catalyst 2 in the presence of ethylene (24) to produce the heterospiro system 234 (91%). Alternatively, a ROM–RCM–CM sequence was employed under similar reaction conditions in the presence of methyl acrylate (232) as a CM partner. The tandem metathesis product 233 was obtained in 68% yield along with the ROM–RCM product 234 in 18% yield (Scheme 47).

[1860-5397-11-199-i47]

Scheme 47: RRM approach to spiro heterocyclic compounds.

Later, 7-azanorbornene 235 has been used in RRM. To this end, compound 235 was subjected to a RRM under the influence of catalyst 3 to deliver the spiro heterocyclic compound 236 (41%). Similarly, compound 237 was treated with catalyst 2 under the same reaction conditions to produce the spiro heterocycle 238 (91%) (Scheme 48).

[1860-5397-11-199-i48]

Scheme 48: RRM approach to spiro heterocyclic compounds.

The RRM strategy has provided an easy access to a variety of 1-azabicyclo[n.3.0]alkenones. For example, when 7-azanorbornene derivative 239 was subjected to a ROM–RCM sequence by treatment with catalyst 2 in toluene in the presence of ethylene (24) delivered the pyrrolizidine system 240 in 63% yield [51]. The regioselective formation of 240 may be attributed to the facile formation of a Ru–carbene intermediate where the metal participates on the side opposite to that of the methyl ester and thereby minimizing the steric crowding between ruthenium and carbonyl oxygen of an ester functionality (Scheme 49). Homologous starting material 242 underwent a RRM with catalyst 2 in the presence of ethylene (24) at 80 °C to produce indolizidine-based compound 243 in 63% yield. Under similar reaction conditions, the azabicyclic system 244 generated pyrrolo[1,2-a]azepine derivative 245. When the RRM protocol was applied to compounds 246a–c with different bridgehead substituents, they also generated the corresponding pyrrolo[1,2-a]azepine derivatives 247a–c in good yields with a high degree of regioselectivity (Scheme 50).

[1860-5397-11-199-i49]

Scheme 49: RRM approach to regioselective pyrrolizidine system 240.

[1860-5397-11-199-i50]

Scheme 50: RRM approach to functionalized bicyclic derivatives.

Treatment of ether-bridged triene 248 with catalyst 2 in chloroform at 50 °C generated the spiroannulated pyrrolidine 249 in 68% yield. However, when the reaction was performed in toluene at 80 °C, the isomeric tricyclic compound 250 was afforded in 34% yield and tricyclic derivative 249 was obtained in 37% yield (Scheme 51).

[1860-5397-11-199-i51]

Scheme 51: RRM approach to tricyclic derivatives 249 and 250.

Rainier’s group [52] has successfully demonstrated the synthesis of various perhydroindolines by adopting a ROM–RCM cascade using catalyst 2 starting with 7-azanorbornene derivative 251. In this context, RRM precursors such as 252 and 253 were obtained from 251 by detosylation sequence. Later, they were subjected to a RRM protocol under the influence of catalyst 2 in the presence of ethylene (24) to generate the expected rearranged products 254 and 255, respectively (Scheme 52 ).

[1860-5397-11-199-i52]

Scheme 52: RRM approach to perhydroindoline derivative and spiro system.

Oxanorbornene systems

Lee and co-workers [53] have successfully constructed a fused bis(oxacyclic) system useful towards the formal total synthesis of dysiherbaine and neodysiherbaine via the RRM protocol. To this end, the oxabicyclo[2.2.1]hept-5-ene 256 was subjected to a RRM cascade with catalyst 2 in dichloromethane to produce pyran derivative 257 in 84% yield, which serves as a core structural unit of disyherbaine. Highly functionalized pyran derivative 259 was obtained by the reaction of 256 with catalyst 5 in the presence of vinyl acetate (258) (Scheme 53).

[1860-5397-11-199-i53]

Scheme 53: RRM approach to bicyclic pyran derivatives.

The RRM approach is useful to design diverse analogs of the marine toxin dysiherbaine, which displays antagonistic activity on ionotropic glutamate receptors from oxanorbornenes [54]. The report reveals the regiochemical directing effect of the exocyclic amidocarbonyl group in a ROM sequence of norbornenes. When the 7-oxanorbornene 260 containing an exocyclic amidocarbonyl moiety was subjected to a metathesis reaction using catalyst 5 in the presence of vinyl acetate (258) at room temperature, the required RRM product 261 was generated in 87% yield with high regio- (>99%) and good stereoselectivity (E/Z = 13:1). Next, tricyclic compound 263 was generated in quantitative yield when the oxanorbornene derivative 262 was subjected to a metathesis with catalyst 5 in the presence of vinyl acetate (258) at room temperature. On the other hand, when the norbornene derivative 264 without the N-benzylaminocarbonyl side chain was subjected to a metathesis under similar reaction conditions a mixture of four products (30:28:6:1) was obtained in 31% combined yield (Scheme 54).

[1860-5397-11-199-i54]

Scheme 54: RRM of various functionalized oxanorbornene systems.

Phelligridin G, a natural product isolated from the fruiting body of P. igniarius, is a well-known anticancer agent. To assemble the spiro-fused furanone core of phelligridin G, Wright and Cooper [55] have used a RRM process as a key step. Wittig olefination of furylbenzaldehyde derivative 265 using methyltriphenylphosphonium bromide in the presence of n-BuLi provided styrylfuran 270 in 72% yield. The DA reaction of styrene derivative 270 with DMAD 129 at 40 °C yielded oxabridged compound 268. Another route to 268 involves a DA reaction of 265 with DMAD at 55 °C for longer reaction time (3 days) and sequential Wittig olefination. The spiro compound 269 was obtained from oxabicyclo adduct 268 by a domino metathesis sequence in the presence of catalyst 2. Moreover, compound 269 was obtained as a single diastereomer and constitutes the core structure of phelligridin G (Scheme 55).

[1860-5397-11-199-i55]

Scheme 55: RRM to assemble the spiro fused-furanone core unit. (i) 129, benzene, 55 °C, 3 days; (ii) Ph3P=CH2Br, n-BuLi, THF, 0 °C; (iii) 129, benzene, 40 °C, 24 h; (iv) catalyst 2 (10 mol %), CH2Cl2, 35 °C.

In 2009, Hanson’s group reported [56] the synthesis of skeletally diverse bi-, and tricyclic sultam derivatives (sulfonamide analogs) using norbornenyl sultam 272 as a core unit assembled by an intramolecular Diels–Alder (IMDA) reaction via a domino ROM–RCM–CM cascade. Diversity has been incorporated by using various cross-metathesis partners (Scheme 56).

[1860-5397-11-199-i56]

Scheme 56: RRM protocol to norbornenyl sultam systems.

Basso and co-workers [57] have demonstrated a tandem Ugi–ROM–RCM protocol towards the synthesis of the 2-aza-7-oxabicyclo[4.3.0]nonane framework by employing catalyst 2. They begin the synthesis with N-allyl-3-endo-amino-7-oxabicyclo[2.2.1]hept-5-ene-2-exo-carboxylic acid (277) and it was used in an Ugi 5-centre-4-component reaction (U-5C-4CR) with a wide variety of aldehydes and isocyanides. Subsequently, the products obtained (e.g., 278) were subjected to a RRM protocol with catalyst 2 to generate the required 2-aza-7-oxabicyclo systems such as 279 (Scheme 57). The advantage of this approach is to provide a simple and short synthetic route to complex polycycles containing the 2-aza-7-oxabicyclo[4.3.0]nonane framework.

[1860-5397-11-199-i57]

Scheme 57: Ugi-RRM protocol for the synthesis of 2-aza-7-oxabicyclo system.

Blanchard and co-workers [58] have reported a novel protocol for the synthesis of spiro- and dispiroketals. The required oxabicyclic derivatives such as 280 were synthesized using α-alkoxyfurans by employing [4 + 2] and/or [4 + 3] cycloaddition reactions. Further, they used a RRM protocol in the presence of catalyst 2 to generate the spiroketal derivative 281 (Scheme 58).

[1860-5397-11-199-i58]

Scheme 58: Synthesis of spiroketal systems via RRM protocol.

Ikoma and co-workers [59] have reported a short synthetic sequence to cis-fused heterocycles by employing the 7-oxanorbornene system 282. In this regard, compound 282 has been prepared by an intramolecular DA reaction as a key step and later, it was subjected to a RRM with catalyst 5 in the presence of ethylene (24) to generate the cis-fused heterotricyclic system 283 (Scheme 59).

[1860-5397-11-199-i59]

Scheme 59: RRM approach to cis-fused heterotricyclic system.

Quinn and co-workers [60] have demonstrated a simple approach to the synthesis of 2,6-dioxabicyclo[3.3.0]octenes 286 starting with the vinyl ether 284 derived from endo-7-oxanorbornene-2-ol by employing a tandem RRM–CM protocol (Scheme 60).

[1860-5397-11-199-i60]

Scheme 60: RRM protocol to functionalized bicyclic systems.

Norbornene systems containing two heteroatoms

Kouklovsky and Vincent have disclosed the RRM of nitroso Diels–Alder (NDA) adducts with a variety of alkenes under microwave or conventional heating conditions by employing catalyst 2 or catalyst 5 to generate various bicyclic compounds [61]. In this regard, compound 287 was subjected to a RRM cascade by employing catalyst 2 in the presence of but-3-en-1-ol (288) under optimized reaction conditions (MW, toluene, 80 °C) and the expected tandem metathesis product 289b was obtained along with the ROM–RCM product 289a. These compounds are useful synthones for the alkaloids synthesis (Scheme 61). In another instance, they also studied the efficiency of this method by isolating the RCM product of the ROM–CM byproduct 290, which was recovered in the ROM–RCM–CM cascade (Scheme 62).

[1860-5397-11-199-i61]

Scheme 61: ROM/RCM/CM cascade to generate bicyclic scaffolds.

[1860-5397-11-199-i62]

Scheme 62: RCM of ROM/CM product.

Kouklovsky and co-workers [62] have described a stereoselective synthesis of 2-(2-hydroxyalkyl)piperidine alkaloids by employing a RRM of NDA adduct 293. The required building block 293 has been prepared via NDA reaction of compound 292 and cyclopentadiene (111). Later, the DA adduct was subjected to a RRM under the influence of catalyst 2 in the presence of but-2-ene (294) to generate the bicyclic isoxazolidine derivative 295. By keeping the bicyclic isoxazolidine ring system intact, this protocol opened an efficient strategy for the formal synthesis of porantheridine and a total synthesis of andrachcinidine (Scheme 63).

[1860-5397-11-199-i63]

Scheme 63: RRM protocol to bicyclic isoxazolidine ring system.

They also reported the formal synthesis of (±)-porantheridine (301) and total synthesis of (±)-8-epihalosaline (300) via a sequential NDA reaction and a RRM [63]. The bicyclic compound 299 was identified as a key building block for the synthesis of 8-epihalosaline (300) and porantheridine (301). To this end, but-3-enoic acid (296) was converted to the required compound 297, which on subjection to NDA in the presence of cyclopentadiene (111) furnished the desired cycloadduct 298 (61% overall yield). Later, it was subjected to the RRM cascade under the influence of catalyst 2 in the presence of 294 to obtain the desired precursor 299 (75% yield, Scheme 64).

[1860-5397-11-199-i64]

Scheme 64: RRM approach toward the total synthesis of (±)-8-epihalosaline (300).

Bicyclo[2.2.2]octene systems

Ghosh and co-workers [40] demonstrated that a RRM approach generates the decalin system 304 rather than the expected 7/6 fused bicyclic system 305. The decalin system has been generated via ROM–RCM starting with bicyclo[2.2.2]octene derivative 303. In this context, compound 302 was initially reacted with catalyst 1 in the presence of ethylene (24) to give 303. Further, treatment with catalyst 2 gave the decalin derivative 304 rather than expected compound 305. However, the metathesis of compound 306, prepared by an independent route produced the expected RCM product 305 in 70% yield (Scheme 65).

[1860-5397-11-199-i65]

Scheme 65: Sequential RRM approach to decalin 304 and 7/6 fused 305 systems.

Kimber and co-workers [64] have described the synthesis of various carbocylic scaffolds by utilizing the RRM protocol involving catalyst 2. They identified the bicyclo[2.2.2]oct-2-en-7-one (307) as the key building block, which was transformed into a mixture of alcohols such as syn-309a and anti-309b in 53% and 24% yield, respectively as a separable diastereomeric mixture (dr, 2:1 ratio). To this end, the syn-product 309a effectively gave the RRM product 310 related to the bicyclo[3.3.1] system with catalyst 2 in the presence of ethylene (24). Alternatively, the anti-product 309b gave the corresponding trans-fused [4.3.0]nonene derivative 311 in 24% yield (Scheme 66).

[1860-5397-11-199-i66]

Scheme 66: RRM protocol to various fused carbocyclic derivatives.

Liao and co-workers [65] have employed the RRM protocol with the DA adduct derived from masked o-benzoquinones (MOBs). Here, they demonstrated an efficient RRM protocol for the synthesis of cis-hydrindenols starting with a readily available starting material such as 2-methoxyphenols. To this end, 2-allylbicyclo[2.2.2]octenol derivative 313 was identified as a key building block in the synthetic sequence, which was prepared from bicyclic system 312 in two steps. When the bicyclic compound 313 (endo isomer) was subjected to a RRM sequence with catalyst 2 in the presence of ethylene (24) at room temperature the desired cis-hydrindenols 315a (95%), 315b (95%) were obtained in excellent yield (Scheme 67).

[1860-5397-11-199-i67]

Scheme 67: RRM to cis-hydrindenol derivatives.

They have also shown that the RRM protocol is applicable with 2-allylbicyclo[2.2.2]octenol derivative 316. The building block 316 required for this purpose has been generated via the DA reaction as a key step starting with 2-methoxyphenol. Later, compound 316 was subjected to a RRM under the influence of catalyst 2 in the presence of ethylene (24) to deliver the expected rearranged product 317 (Scheme 68).

[1860-5397-11-199-i68]

Scheme 68: RRM protocol towards the cis-hydrindenol derivatives.

Vanderwal and co-workers [66] described the synthesis of polycyclic lactams obtained by arene/allene cycloaddition, discovered by Himbert and Henn were found to undergo a RRM in a facile manner in the presence of catalyst 6 to produce complex polycyclic lactams. In this regard, the required building block 319 was obtained from compound 318 by cycloaddition reaction. A variety of complex molecular frames were accessed via the RRM sequence under the influence of catalyst 6 in toluene at 50–100 °C in the presence of 24. The procedure is suitable for the preparation of diverse polycyclic lactams with a variety of substitution patterns (Scheme 69).

[1860-5397-11-199-i69]

Scheme 69: RRM approach toward the synthesis of diversed polycyclic lactams.

Kotha and Ravikumar [44] have successfully executed the RRM protocol for the synthesis of condensed polycyclic systems. To this end, bicyclo[2.2.2]octene derivative 321 has been identified as a key starting material. The required key building block 323 has been prepared from the known bis-DA adduct 321 [67] via allyl Grignard addition followed by O-allylation sequence. The starting cycloadduct 321 was obtained by the double DA reaction between 1,3-cyclohexadiene and 1,4-benzoquinone. Further, treatment of 323 with catalyst 2 in the presence of titanium isopropoxide furnished the expected RRM product 324 in 92% yield (Scheme 70).

[1860-5397-11-199-i70]

Scheme 70: RRM approach towards synthesis of hexacyclic compound 324.

Bicyclo[2.2.2]octene systems containing nitrogen

To design lycopodium alkaloids, Barbe and co-workers [68] have used RRM judiciously. The required precursor 326 suitable for RRM has been prepared from pyridine (325) in four steps on gram scale. Later, the azabicyclic system was reacted with catalyst 2 to generate the desired hydroquinoline derivative 327 in 81% yield. Further, they have used the bicyclic compound 327 as a key building block in the total synthesis of (+)-luciduline (Scheme 71).

[1860-5397-11-199-i71]

Scheme 71: RRM protocol to generate luciduline precursor 327 with catalyst 2.

Lepadins are natural products consisting of cis-fused decahydroquinoline subunits and they display cytotoxic activity against many human cancer cell lines. The total synthesis of (+)-lepadin B developed by Charette and Barbe [69] utilized a RCM–ROM as key step. In this regard the azabicyclic system 329 (obtained from pyridine (325)) was subjected to a RRM sequence by employing catalyst 2 at 80 °C in toluene to furnish the rearranged product 330 (79%). Further, the building block 330 was used in the stereoselective total synthesis of lepadin B (Scheme 72).

[1860-5397-11-199-i72]

Scheme 72: RRM protocol to key building block 330.

Bicyco[3.2.1]octene derivatives

Norhalichondrin B is a marine polyether belonging to the halichondrin family and its macrolactone analog has displayed anticancer activity. Phillips and co-workers [70] have described a total synthesis of norhalichondrin B in 37 steps from β-furylethanol. Interesting feature of this synthetic sequence is the tactical utilization of tandem ROM–RCM protocol towards the synthesis of the key intermediate 335. In this reaction, the required RRM precursor 333 was obtained from diazo ester 331 in five steps. Further, the RRM of 333 with catalyst 2 furnished the required pyran derivative 334 (71%). Next, the fused ether 334 was transformed into the desired intermediate 335 in eight steps, which is a key intermediate required for the synthesis of norhalichondrin B (Scheme 73).

[1860-5397-11-199-i73]

Scheme 73: RRM approach towards the synthesis of key intermediate 335.

To expand the scope of the RRM methodology, Wright and Cooper [55] reported the synthesis of a highly functionalized pyran system by employing a RRM as a key step. To this end, 2-phenylfuran derivatives 265 and 270 were reacted with tetrachlorocyclopropene (TCCP, 336) followed by olefination to result the required oxabicyclo[3.2.1]octene derivative 338. Later, the RRM of the styrene derivative 338 with catalyst 2 delivered a highly-functionalized spiro-pyran derivative 339 in 48% yield (Scheme 74).

[1860-5397-11-199-i74]

Scheme 74: RRM protocol to highly functionalized spiro-pyran system 339.

The Dysiherbaine and acetogenin groups of natural products have been synthesized by the RRM approach. In this regard, secondary alcohol derivatives related to 8-oxabicyclo[3.2.1]octenes such as 341a,b,c were used as potential precursors for the synthesis of a variety of cyclic polyethers [71]. Allylation of 340a–c using sodium hydride and allyl bromide (37) in the presence of a phase-transfer catalyst such as tetrabutylammonium iodide generated bicyclic compounds 341a–c. The RRM of these ether derivatives 341a–c was performed under ethylene (24) atmosphere with catalyst 5 to generate the dihydrofuran derivatives 342a–c. When compounds 340d, 340e and 340f were subjected to a metathesis protocol by treatment with catalyst 2 under ethylene (24) atmosphere in the presence of 1,4-benzoquinone, cis-fused hexahydrofuro[3,2-b]pyran core containing compounds 343d, 343e and 343f were obtained via RRM in good yields (50–75%) (Scheme 75).

[1860-5397-11-199-i75]

Scheme 75: RRM to various bicyclic polyether derivatives.

Conclusion

RRM involving ROM–RCM under the influence of various Ru–carbene complexes in one-pot sequence generate various complex targets. It is an atom economic process producing a wide range of polycyclic compounds containing highly demanding structures efficiently. Starting with relatively simple substrates, the final compounds obtained by the RRM process are generally difficult to synthesize by conventional synthetic routes. Various examples described here have clearly established the power and scope of this methodology. We believe that an increasing number of natural as well as non-natural products of high structural complexity have assembled by the RRM process and this activity will continue with more vigour in the future.

Acknowledgements

We thank CSIR and the Department of Science and Technology (DST), New Delhi for the financial support. SK thanks DST for the award of a J. C. Bose fellowship. PK thanks DST for Fast Track Research Grant.

References

  1. Zuercher, W. J.; Hashimoto, M.; Grubbs, R. H. J. Am. Chem. Soc. 1996, 118, 6634–6640. doi:10.1021/ja9606743
    Return to citation in text: [1]
  2. Harrity, J. P. A.; Visser, M. S.; Gleason, J. D.; Hoveyda, A. H. J. Am. Chem. Soc. 1997, 119, 1488–1489. doi:10.1021/ja9636597
    Return to citation in text: [1]
  3. Holub, N.; Blechert, S. Chem. – Asian J. 2007, 2, 1064–1082. doi:10.1002/asia.200700072
    Return to citation in text: [1]
  4. Nolan, S. P.; Clavier, H. Chem. Soc. Rev. 2010, 39, 3305–3316. doi:10.1039/b912410c
    Return to citation in text: [1]
  5. Schmidt, B.; Krehl, S. Domino and Other Olefin Metathesis Reaction Sequences. In Olefin Metathesis: Theory and Practice; Grela, K., Ed.; John Wiley & Sons: Hoboken, 2014; pp 187–232.
    Return to citation in text: [1]
  6. Grubbs, R. H.; O’Leary, D. J. Handbook of Metathesis, 2nd ed.; Applications in Organic Synthesis, Vol. 2; Wiley-VCH: Weinheim, Germany, 2015.
    Return to citation in text: [1]
  7. Zhu, Z.-B.; Shi, M. Org. Lett. 2010, 12, 4462–4465. doi:10.1021/ol101455c
    Return to citation in text: [1]
  8. Miege, F.; Meyer, C.; Cossy, J. Org. Lett. 2010, 12, 248–251. doi:10.1021/ol9025606
    Return to citation in text: [1]
  9. Maougal, E.; Dalençon, S.; Pearson-Long, M. S. M.; Mathé-Allainmat, M.; Lebreton, J.; Legoupy, S. Synthesis 2014, 46, 3268–3272. doi:10.1055/s-0034-1378663
    Return to citation in text: [1]
  10. White, B. H.; Snapper, M. L. J. Am. Chem. Soc. 2003, 125, 14901–14904. doi:10.1021/ja037656n
    Return to citation in text: [1]
  11. Zhang, F.; Simpkins, N. S.; Blake, A. J. Org. Biomol. Chem. 2009, 7, 1963–1979. doi:10.1039/b900189A
    Return to citation in text: [1]
  12. Pandya, B. A.; Snapper, M. L. J. Org. Chem. 2008, 73, 3754–3758. doi:10.1021/jo702702s
    Return to citation in text: [1]
  13. Tang, B.; Bray, C. D.; Pattenden, G.; Rogers, J. Tetrahedron 2010, 66, 2492–2500. doi:10.1016/j.tet.2010.01.059
    Return to citation in text: [1]
  14. Han, J.-c.; Li, F.; Li, C.-c. J. Am. Chem. Soc. 2014, 136, 13610–13613. doi:10.1021/ja5084927
    Return to citation in text: [1]
  15. Halle, M. B.; Fernandes, R. A. RSC Adv. 2014, 4, 63342–63348. doi:10.1039/C4RA10937F
    Return to citation in text: [1]
  16. Kress, S.; Weckesser, J.; Schulz, S. R.; Blechert, S. Eur. J. Org. Chem. 2013, 1346–1355. doi:10.1002/ejoc.201201516
    Return to citation in text: [1]
  17. Li, J.; Lee, D. Chem. Sci. 2012, 3, 3296–3301. doi:10.1039/C2SC20812A
    Return to citation in text: [1]
  18. Ma, S.; Ni, B. Chem. – Eur. J. 2004, 10, 3286–3300. doi:10.1002/chem.200305581
    Return to citation in text: [1]
  19. Groaz, E.; Banti, D.; North, M. Tetrahedron 2008, 64, 204–218. doi:10.1016/j.tet.2007.10.076
    Return to citation in text: [1] [2]
  20. Gao, F.; Stamp, C. T. M.; Thornton, P. D.; Cameron, T. S.; Doyle, L. E.; Miller, D. O.; Burnell, D. J. Chem. Commun. 2012, 48, 233–235. doi:10.1039/C1CC15452D
    Return to citation in text: [1]
  21. Donnard, M.; Tschamber, T.; Le Nouën, D.; Desrat, S.; Hinsinger, K.; Eustache, J. Tetrahedron 2011, 67, 339–357. doi:10.1016/j.tet.2010.11.036
    Return to citation in text: [1]
  22. Donnard, M.; Tschamber, T.; Desrat, S.; Hinsinger, K.; Eustache, J. Tetrahedron Lett. 2008, 49, 1192–1195. doi:10.1016/j.tetlet.2007.12.047
    Return to citation in text: [1]
  23. Donnard, M.; Tschamber, T.; Eustache, J. Tetrahedron Lett. 2008, 49, 7325–7327. doi:10.1016/j.tetlet.2008.10.048
    Return to citation in text: [1]
  24. Mori, M.; Kuzuba, Y.; Kitamura, T.; Sato, Y. Org. Lett. 2002, 4, 3855–3858. doi:10.1021/ol026696d
    Return to citation in text: [1]
  25. Holtsclaw, J.; Koreeda, M. Org. Lett. 2004, 6, 3719–3722. doi:10.1021/ol048650l
    Return to citation in text: [1]
  26. Wrobleski, A.; Sahasrabudhe, K.; Aubé, J. J. Am. Chem. Soc. 2002, 124, 9974–9975. doi:10.1021/ja027113y
    Return to citation in text: [1]
  27. Henderson, J. A.; Phillips, A. J. Angew. Chem., Int. Ed. 2008, 47, 8499–8501. doi:10.1002/anie.200803593
    Return to citation in text: [1]
  28. Shibatomi, K.; Kobayashi, F.; Narayama, A.; Fujisawa, I.; Iwasa, S. Chem. Commun. 2012, 48, 413–415. doi:10.1039/C1CC15889A
    Return to citation in text: [1]
  29. Funel, J.-A.; Prunet, J. Synlett 2005, 235–238. doi:10.1055/s-2004-837200
    Return to citation in text: [1]
  30. Kotha, S.; Ravikumar, O. Eur. J. Org. Chem. 2014, 5582–5590. doi:10.1002/ejoc.201402273
    Return to citation in text: [1]
  31. Kotha, S.; Ravikumar, O. Tetrahedron Lett. 2014, 55, 5781–5784. doi:10.1016/j.tetlet.2014.08.108
    Return to citation in text: [1]
  32. Kotha, S.; Gunta, R. Beilstein J. Org. Chem. 2015, 11, 1727–1731. doi:10.3762/bjoc.11.188
    Return to citation in text: [1]
  33. Kotha, S.; Gunta, R. Beilstein J. Org. Chem. 2015, 11, 1373–1378. doi:10.3762/bjoc.11.148
    Return to citation in text: [1]
  34. Kotha, S.; Ravikumar, O.; Majhi, J. Beilstein J. Org. Chem. 2015, 11, 1503–1508. doi:10.3762/bjoc.11.163
    Return to citation in text: [1]
  35. Maity, S.; Ghosh, S. Tetrahedron Lett. 2008, 49, 1133–1136. doi:10.1016/j.tetlet.2007.12.064
    Return to citation in text: [1]
  36. Maity, S.; Ghosh, S. Tetrahedron 2009, 65, 9202–9210. doi:10.1016/j.tet.2009.09.029
    Return to citation in text: [1]
  37. Mondal, S.; Malik, C. K.; Ghosh, S. Tetrahedron Lett. 2008, 49, 5649–5651. doi:10.1016/j.tetlet.2008.07.083
    Return to citation in text: [1]
  38. Mondal, S.; Yadav, R. N.; Ghosh, S. Tetrahedron Lett. 2009, 50, 5277–5279. doi:10.1016/j.tetlet.2009.07.012
    Return to citation in text: [1]
  39. Matcha, K.; Maity, S.; Malik, C. K.; Ghosh, S. Tetrahedron Lett. 2010, 51, 2754–2757. doi:10.1016/j.tetlet.2010.03.074
    Return to citation in text: [1]
  40. Bose, S.; Ghosh, M.; Ghosh, S. J. Org. Chem. 2012, 77, 6345–6350. doi:10.1021/jo300945b
    Return to citation in text: [1] [2]
  41. Malik, C. K.; Yadav, R. N.; Drew, M. G. B.; Ghosh, S. J. Org. Chem. 2009, 74, 1957–1963. doi:10.1021/jo802077t
    Return to citation in text: [1]
  42. Malik, C. K.; Ghosh, S. Org. Lett. 2007, 9, 2537–2540. doi:10.1021/ol070906a
    Return to citation in text: [1]
  43. Datta, R.; Bose, S.; Viththlbhai, P. B.; Ghosh, S. Tetrahedron Lett. 2014, 55, 3538–3540. doi:10.1016/j.tetlet.2014.04.091
    Return to citation in text: [1]
  44. Kotha, S.; Ravikumar, O. Beilstein J. Org. Chem. 2015, 11, 1259–1264. doi:10.3762/bjoc.11.140
    Return to citation in text: [1] [2]
  45. Higashibayashi, S.; Tsuruoka, R.; Soujanya, Y.; Purushotham, U.; Sastry, G. N.; Seki, S.; Ishikawa, T.; Toyota, S.; Sakurai, H. Bull. Chem. Soc. Jpn. 2012, 85, 450–467. doi:10.1246/bcsj.20110286
    Return to citation in text: [1]
  46. Nguyen, N. N. M.; Leclere, M.; Stogaitis, N.; Fallis, A. G. Org. Lett. 2010, 12, 1684–1687. doi:10.1021/ol100150f
    Return to citation in text: [1]
  47. Thomas, G. L.; Spandl, R. J.; Glansdorp, F. G.; Welch, M.; Bender, A.; Cockfield, J.; Lindsay, J. A.; Bryant, C.; Brown, D. F. J.; Loiseleur, O.; Rudyk, H.; Ladlow, M.; Spring, D. R. Angew. Chem., Int. Ed. 2008, 47, 2808–2812. doi:10.1002/anie.200705415
    Return to citation in text: [1]
  48. Standen, P. E.; Dodia, D.; Elsegood, M. R. J.; Teat, S. J.; Kimber, M. C. Org. Biomol. Chem. 2012, 10, 8669–8676. doi:10.1039/c2ob26784e
    Return to citation in text: [1]
  49. Miura, Y.; Hayashi, N.; Yokoshima, S.; Fukuyama, T. J. Am. Chem. Soc. 2012, 134, 11995–11997. doi:10.1021/ja305856q
    Return to citation in text: [1]
  50. Carreras, J.; Avenoza, A.; Busto, J. H.; Peregrina, J. M. J. Org. Chem. 2011, 76, 3381–3391. doi:10.1021/jo200321t
    Return to citation in text: [1]
  51. Rojas, V.; Carreras, J.; Avenoza, A.; Busto, J. H.; Peregrina, J. M. Eur. J. Org. Chem. 2013, 3817–3824. doi:10.1002/ejoc.201300126
    Return to citation in text: [1]
  52. Liu, Z.; Rainier, J. D. Org. Lett. 2006, 8, 459–462. doi:10.1021/ol052741g
    Return to citation in text: [1]
  53. Lee, H.-Y.; Lee, S.-S.; Kim, H. S.; Lee, K. M. Eur. J. Org. Chem. 2012, 4192–4199. doi:10.1002/ejoc.201200439
    Return to citation in text: [1]
  54. Ikoma, M.; Oikawa, M.; Gill, M. B.; Swanson, G. T.; Sakai, R.; Shimamoto, K.; Sasaki, M. Eur. J. Org. Chem. 2008, 5215–5220. doi:10.1002/ejoc.200800704
    Return to citation in text: [1]
  55. Cooper, H. D.; Wright, D. L. Molecules 2013, 18, 2438–2448. doi:10.3390/molecules18022438
    Return to citation in text: [1] [2]
  56. Jeon, K. O.; Rayabarapu, D.; Rolfe, A.; Volp, K.; Omar, I.; Hanson, P. R. Tetrahedron 2009, 65, 4992–5000. doi:10.1016/j.tet.2009.03.080
    Return to citation in text: [1]
  57. Sonaglia, L.; Banfi, L.; Riva, R.; Basso, A. Tetrahedron Lett. 2012, 53, 6516–6518. doi:10.1016/j.tetlet.2012.09.076
    Return to citation in text: [1]
  58. Mandel, J.; Dubois, N.; Neuburger, M.; Blanchard, N. Chem. Commun. 2011, 47, 10284–10286. doi:10.1039/C1CC14329H
    Return to citation in text: [1]
  59. Ikoma, M.; Oikawa, M.; Sasaki, M. Tetrahedron 2008, 64, 2740–2749. doi:10.1016/j.tet.2008.01.067
    Return to citation in text: [1]
  60. Quinn, K. J.; Curto, J. M.; Faherty, E. E.; Cammarano, C. M. Tetrahedron Lett. 2008, 49, 5238–5240. doi:10.1016/j.tetlet.2008.06.115
    Return to citation in text: [1]
  61. Vincent, G.; Kouklovsky, C. Chem. – Eur. J. 2011, 17, 2972–2980. doi:10.1002/chem.201002558
    Return to citation in text: [1]
  62. Vincent, G.; Karila, D.; Khalil, G.; Sancibrao, P.; Gori, D.; Kouklovsky, C. Chem. – Eur. J. 2013, 19, 9358–9365. doi:10.1002/chem.201300836
    Return to citation in text: [1]
  63. Sancibrao, P.; Karila, D.; Kouklovsky, C.; Vincent, G. J. Org. Chem. 2010, 75, 4333–4336. doi:10.1021/jo100768d
    Return to citation in text: [1]
  64. Standen, P. E.; Kimber, M. C. Tetrahedron Lett. 2013, 54, 4098–4101. doi:10.1016/j.tetlet.2013.05.112
    Return to citation in text: [1]
  65. Tsao, K.-W.; Devendar, B.; Liao, C.-C. Tetrahedron Lett. 2013, 54, 3055–3059. doi:10.1016/j.tetlet.2013.03.142
    Return to citation in text: [1]
  66. Lam, J. K.; Schmidt, Y.; Vanderwal, C. D. Org. Lett. 2012, 14, 5566–5569. doi:10.1021/ol302680m
    Return to citation in text: [1]
  67. Valiulin, R. A.; Arisco, T. M.; Kutateladze, A. G. Org. Lett. 2010, 12, 3398–3401. doi:10.1021/ol101297b
    Return to citation in text: [1]
  68. Barbe, G.; Fiset, D.; Charette, A. B. J. Org. Chem. 2011, 76, 5354–5362. doi:10.1021/jo200745n
    Return to citation in text: [1]
  69. Barbe, G.; Charette, A. B. J. Am. Chem. Soc. 2008, 130, 13873–13875. doi:10.1021/ja8068215
    Return to citation in text: [1]
  70. Jackson, K. L.; Henderson, J. A.; Motoyoshi, H.; Phillips, A. J. Angew. Chem., Int. Ed. 2009, 48, 2346–2350. doi:10.1002/anie.200806111
    Return to citation in text: [1]
  71. Escavabaja, P.; Viala, J.; Coquerel, Y.; Rodriguez, J. Adv. Synth. Catal. 2012, 354, 3200–3204. doi:10.1002/adsc.201200665
    Return to citation in text: [1]

© 2015 Kotha et al; licensee Beilstein-Institut.
This is an Open Access article under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
The license is subject to the Beilstein Journal of Organic Chemistry terms and conditions: (http://www.beilstein-journals.org/bjoc)

 
Back to Article List