2
Alkene Metathesis in Total Synthesis
2.1
Total Syntheses Based on a Ring-Closing-Metathesis Reaction
The ring-closing metathesis (RCM) reaction is one of the most frequently employed
methodologies to construct cyclic systems, becoming a common synthetic tool in the
modern organic synthetic laboratory.
Indeed, a practically limitless array of carbocyclic or heterocyclic ring systems,
including small-, medium-, and large-sized rings, can be constructed utilizing this
transformation. In the context of total synthesis, the preparation of small- and medium-sized
rings usually plays a complementary role implemented along the synthetic path. However,
in the case of macrocyclic systems, the RCM reaction commonly constitutes a crucial
step for the completion of the synthesis. For this reason, in this section, we would
like to highlight representative examples reported in the last few years in which
the RCM reaction was the key step. As a first series of examples in this category,
it is worth mentioning the polyketide-type compounds ripostatin B (28)[35] and A (29),[36] FD-895 (30),[37] cruentaren A (31),[38] pikromycin (32),[39] zampanolide (33),[40] amphidinolide G (34),[41] palmerolide A (35),[42] ecklonialactone B (36),[43] gambieric acid A (37),[44] nominal gobienine A (38),[45] trienomycin A (39) and F (40),[46] 13-demethyllyngbyasolide B (41),[47] incednam (42),[48] sekothrixide (43),[49] aspicillin (44),[50] macrolide fragment of FD-891 (45),[51] carolacton (46)[52] (Figure [2]), cytospolide P (47),[53] pectenotoxin (48),[54] Sch725674 (49),[55] aspergillide B (revised structure, 50),[56] iriometolide 3a (51),[57] paecilomycin B (52),[58] fidaxomicin (53),[59] exiguolide (54),[60] neopeltolide (55),[61] methynolide (56),[62] and iriomoteolide-2a (57)[63] (Figure [3]). In addition, the group of alkaloids and cyclodepsipeptides are illustrated with
the examples of manzamine A (58),[64] isoschizogamine (59),[65] vertine (60),[66] haliclamide (61),[67] 5-epi-torrubiellutin (62),[68] petrosin (63),[69] the YM-254890 analogue 64,[70] nannocystin A (65) and A0 (66),[71] nakadomarin A (67),[72] marineosin A (68),[73] kanamienamide (69),[74] and 15-epi-aetheramide A (70)[75] (Figure [4]).
Figure 1 Common catalysts used for alkene- and enyne-metathesis reactions: evolution of their
design and development
Finally, among the total syntheses of terpenes that employ the RCM reaction, particularly
noteworthy are the syntheses of 17-deoxyprovidencin (71),[76] uprolide G acetate (72),[77] terreumol C (73),[78] pavidolide B (74),[79] and boscartin F (75)[80] (Figure [5]). As an indication of the effectiveness of the RCM processes, the yields achieved
in many cases were reasonable to excellent (44–99% yields), with some exceptions with
lower yields (17–34%), and in general excellent stereoselectivities, with some exceptions
[ecklonialactone B (36),[43] fidaxomicin (53),[59] iriomoteolide-2a (57),[63] nakadomarin A (67)[72]], for which the isomeric ratios are indicated.
In addition to the previous contributions, we can find in the literature other cases
that deserve a more detailed description. A first example is the synthesis by Brimble
and co-workers of palmyrolide A (76) (Scheme [2]),[81] a macrolide isolated from Leptotyngbya cf. sp. and Oscillatoria sp. that displays interesting neuroprotective properties, combined with low cytotoxicity.
The synthesis was initially based on a RCM reaction of a starting enamide precursor;
however, this RCM attempt failed despite screening of a wide variety of reaction conditions.
As a consequence of these disappointing results, they considered a sequential RCM/olefin
isomerization from the diolefin 77, with the possibility that the desired isomerization may occur in the presence of
the corresponding Grubbs catalyst. However, despite the RCM reaction of diolefin 77 proceeding efficiently when treated with the Grubbs II catalyst (4) to give macrocycle 78, the isomerization was not observed. This result led them to force the isomerization
by subsequent treatment of the RCM product with RuH(PPh3)3(CO)Cl, which then afforded the natural product 76 in a good overall yield. This synthetic study constituted the first reported isomerization
of an N-allylated tertiary amide in a macrocyclic setting and, furthermore, allowed the revision
of the structure of the natural product, which was erroneously assigned to the initial
proposed structures 79 and 80 (Scheme [2]).
A second relevant example is from Carreira and co-workers, who in a stereodivergent-oriented
approach, described the synthesis of the four diastereomers of Δ9-tetrahydrocannabinol (81) via a RCM reaction from the acyclic precursors 82 (Scheme [3]).[82] Interestingly, the four stereoisomers of 82 were efficiently prepared by an elegant and novel dual catalytic process based on
the use of a set of two chiral catalysts [Ir/(P, olefin) and a secondary amine] for
the enantioselective allylation of 5-methylhex-5-enal. The RCM reaction of each stereoisomer
of 82 was achieved under very mild conditions (25 °C) with the Grubbs II catalyst (4) to provide the corresponding THC precursors 83 in excellent 85–92% yields, which were transformed into the final products in three
additional steps (Scheme [3]). Previous syntheses of Δ9-THC 81 and related cannabinoid-type compounds have also employed the RCM reaction to construct
the cyclohexene ring containing in these intriguing natural products.[83]
Figure 2 Selected macrocyclic polyketides synthesized via a RCM reaction (yields refer only
to the RCM reaction)
Figure 3 Selected macrocyclic polyketides synthesized via a RCM reaction (yields refer only
to the RCM reaction)
Scheme 2 Total synthesis of palmyrolide A
Scheme 3 Total synthesis of Δ9-THC and its stereoisomers
Figure 4 Selected alkaloids and cyclodepsipeptides synthesized via a RCM reaction (yields
refer only to the RCM reaction)
The construction of medium-sized rings containing cyclobutanes represents a challenging
ring-closing metathesis due to the highly strained character of the resulting bicyclic
system. Despite this potential difficulty, the RCM reaction has proven to be highly
efficient in many cases. An instructive example of this is demonstrated in the synthesis
of the polycyclic xanthone-type antibiotic IB-00208 (84), reported by Martin and co-workers (Scheme [4]).[84] Thus, the synthesis of the key cyclobutenone 86 was successfully achieved via a RCM reaction of precursor 85 when subjected to the Grubbs II catalyst (4). The resulting cyclobutenone was rearranged to the desired xanthone 87 under thermal conditions and represents a general route to construct polycyclic benzoquinones.
A few more steps then converted the xanthone 87 into the targeted natural product 84.
As another example of the use of the RCM reaction to fashion cyclobutane-containing
polycyclic systems, we highlight the total synthesis of the cytotoxic meroterpenoid
psiguadial B (88) by Reisman and co-workers (Scheme [5]),[85] in which the RCM reaction of diolefinic cyclobutane 89, carried out with the H-G II catalyst (6) in the presence of 1,4-benzoquinone, afforded in an excellent 93% yield the corresponding
tricyclic derivative 90. After a formidable chemoselective hydrogenation of 90, which was accomplished in an excellent 90% yield by the action of Crabtree’s catalyst,
the resulting product 91 was carried towards the natural product in four additional steps.
A stunning example of the power of the RCM reaction can be found in the construction
of strained ring systems via a transannular process as applied in the recent synthesis
of the cembranolides sarcophytonolide H (92) and isosarcophytonolide D (93) by Takamura and co-workers (Scheme [6]).[86] These natural products, isolated from the soft coral of the genus Sarcophyton, exhibit potent antifouling activities against the larval settlement of barnacle
Balanus amphirite with EC50 values in the low μg/mL range. In these syntheses, the construction of the butenolide
contained in the macrocyclic framework, which is a structural feature of this class
of diterpenes, was achieved via a RCM reaction of the corresponding precursors 94 and 96 by treatment with the H-G II catalyst (6) in toluene at 100 °C. The efficiency of these processes is quite remarkable, given
both the steric encumbrance around the ring-closure site and the complexity of the
system. The completion of the synthesis of both natural products was carried out in
three additional steps from the RCM products.
Figure 5 Selected terpenes synthesized via a RCM reaction (yields refer only to the RCM reaction)
Scheme 4 Total synthesis of IB-00208
Scheme 5 Total synthesis of psiguadial B
Scheme 6 Total synthesis of sarcophytonolide H and isosarcophytonolide D
Further evidence for the utility of the RCM reaction in the construction of sterically
crowded and highly functionalized ring systems is provided by the recent total synthesis
of the highly oxidized diterpenoids ryanodol (97) and related diterpenes, natural products that display powerful pharmacological and
insecticidal activities via their interactions with the ryanodine receptors.[87] The synthesis of the C-ring of ryanodol was undertaken by Inoue and co-workers via
the RCM reaction of the precursor 99 (Scheme [7]),[88] prepared from the allylic diketone 98 in three steps, in 58% yield over four steps, to obtain advanced precursor 100. Remarkably, once again, the presence of unprotected functionalities was well tolerated
under these ruthenium-catalyzed reaction conditions, thus underscoring the mildness
and efficiency of these reactions in a complex setting. From compound 100, the completion of the synthesis of ryanodol (97) was accomplished in ten additional steps with minimal difficulty.
Within the stereoselective RCM processes, the use of chiral catalysts, such as the
catalysts 15–22 designed by Hoveyda, Schrock, and co-workers (see Figure [1]), allows for the preparation of enantiomerically enriched products.[89] Their applications and level of enantioselective control is nicely illustrated in
the synthesis of the alkaloid deoxoapodine (101), a member of a family of alkaloids that share the pentacyclic aspidosperma core.
This asymmetric synthesis was devised by Movassaghi and co-workers based on a desymmetrization
process of the achiral triolefin 102 (Scheme [8]).[90] Thus, achiral compound 102 was subjected to the catalytic activity of the chiral molybdenum pyrrolide complex
17, which proved to be efficient in terms of chemical yield and stereoselectivity in
the synthesis of quebrachamine,[19c] however, in this case, the catalyst offered only moderate enantioselectivity [er
(–)-103/(+)-103 82:18]. This stereochemical outcome was likely influenced by the presence of the
amide group in the starting precursor, which could coordinate with the transition
metal resulting in a consequent reduction of the catalytic activity. In light of these
results, an exploration of other chiral catalysts led to the observation that as the
size of the halide substituent on the complex increased, the level of enantioselectivity
was improved, with a (–)-103/(+)-103 93:7 ratio of enantiomers being achieved when the diiodo complex 19 was employed. As the natural product possessed the opposite configuration, this result
was extended to the enantiomer catalyst ent-19 to obtain the product 103 with the configuration of the natural product in 92% yield and with an enantiomeric
ratio of (+)-103/(–)-103 94:6. With the key product (+)-103 in hand, the synthetic route to the natural product was delineated through the pentacyclic
derivative 104, which was prepared in an efficient manner from (+)-103 according to the steps indicated in Scheme [8]. From pentacycle 104, seven additional steps were required for completion of the synthesis of deoxoapodine
(101).
Scheme 7 Total synthesis of ryanodol
Scheme 8 Total synthesis of deoxoapodine
Scheme 9 Total synthesis of pseudotabersonine
Despite the excellent functional group tolerance and the low sensitivity to steric
hindrances displayed by these catalysts, we can find in the literature numerous examples
in which steric factors appear to play a crucial role in unsuccessful ring-closure
reactions. This is the case for compound 105, which was devised as a potential precursor for the synthesis of the natural product
pseudotabersonine (106) by Martin and co-workers (Scheme [9]).[91] Thus, when 105 was treated with different ruthenium catalysts, including Grubbs I (3), Grubbs II (4), H-G II (6), or even the Grubbs–Stewart catalyst (12), which is especially reactive towards sterically hindered olefins, the desired product
107, resulting from a double RCM process was not observed, instead obtaining the monocyclized
product 108 in moderate yields, together with degradation products. They explored structural
modifications of the precursor, by increasing its reactivity by removing the ester
group, together with a complete optimization study of the RCM reaction. This extensive
study led them to define a set of suitable structural requisites and reaction conditions
to enable the targeted product. Accordingly, having established 109 as a suitable precursor, when this compound was treated with H-G II catalyst (6) in toluene at 100 °C, the desired metathesis product 110 was obtained in 90% yield albeit as an inseparable mixture of cis/trans isomers in 7:10 ratio. After transformation of the diastereomeric mixture 110 into the reduced product 111, the isomers could be separated, and then the required cis isomer was taken forward to completion of the natural product in three more steps.
A noteworthy case in point is the natural product soraphen A1α (112), whose biological activities, including antimicrobial, antifungus, antidiabetes,
and, more recently, anticancer properties, has prompted great interest in chemical
and biological circles. Preliminary synthetic efforts conducted by Ciufolini and co-workers,[92] towards this compound based on RCM were met with failure when they attempted to
effect a ring closure with various acyclic precursors 113–116, different catalysts and several reaction conditions to obtain the corresponding
coveted macrocycles 117–120 (Scheme [10]). Reasoning that steric factors were responsible for the failed RCM reactions, other
investigators attempted the metathesis macrocyclization at less sterically encumbered
sites. To this aim, Kalesse and co-workers[93] explored the RCM reaction of precursor 121 by treatment with different ruthenium-based catalysts, however the results were similarly
unfruitful in all cases (Scheme [10]). In contrast to all these discouraging results, the RCM reaction of the desmethoxy
precursor 123, explored by Micalizio and co-workers (Scheme [10]),[94] provided in high yield (90%) the corresponding macrocycle 124 with Grubbs II catalyst (4), proof that steric factors were responsible for the previous failures. From compound
124, they completed the synthesis of the soraphen analogue 125. Interestingly, this soraphen analogue displayed promising cytotoxic activities against
B-lymphoma cell lines, revealing that the methoxy group present in the natural product
was not essential for biological activity.
Scheme 10 Synthetic attempts towards soraphen A1α via RCM and synthesis of the C11-desmethoxy analogue
A paradigmatic example in which the RCM reaction proved to be completely useless was
in the synthesis of the bryostatins. The well-known fascinating synthesis of the bryostatins
by Trost and co-workers was preceded by an initial attempt of its total synthesis
via a RCM reaction of the diolefins 126 and 127 (Scheme [11]).[95] However, after extensive screening of different catalysts and conditions, they were
unable to obtain the corresponding macrocycles 128 and 129. Furthermore, the implementation of the relay RCM (RRCM) strategy, as we will describe
later, did not provide the expected macrocyclic olefin, when the RRCM precursor 130 was subjected to different ruthenium-based catalysts, providing instead the macrocyclic
product 131 as a result of the incorporation of the extended olefin onto the final cyclization
product. Having ascribed the importance of steric hindrance, imposed by the presence
of the gem-dimethyl system at the allylic position, as the reason for these disappointing results,
in 2017 Thomas and co-workers studied the RCM reaction in model systems of the bryostatins
in which the gem-dimethyl group was removed (Scheme [11]).[96] In their study, the RCM reaction of model compound 132 by exposure to the Grubbs II catalyst (4) furnished the macrocyclic compound 133, albeit in a poor 17% yield (Scheme [11]). These results are an indication of the serious hurdles that this class of systems
presents for the metathesis reaction.
Scheme 11 Synthetic studies on bryostatins via RCM reactions
A recent 2018 example in which a RCM reaction was combined with a final CM reaction
to append the side chain in a stereoselective manner is the synthesis of the halogenated
marine metabolites chlorofucins 134 and 135 and bromofucins 136 and 137 by Paton, Kim, and co-workers (Scheme [12]).[97] These natural products belong to the family of the bioactive acetogenins isolated
from the Laurencia red algae that have generated great synthetic interest, particularly in the development
of efficient methodologies for the regio- and stereocontrolled construction of the
medium-ring oxacycle that characterize these appealing natural products.[98] Furthermore, for the chlorofucins 134 and 135 and bromofucins 136 and 137, their absolute configurations were not established so their asymmetric syntheses
could unambiguously confirm their proposed configurations. Thus, the formation of
the common oxocene derivative 139 was efficiently undertaken by exposure of the acyclic precursor 138 to Grubbs II catalyst (4). For the incorporation of the halogen with concomitant intramolecular etherification,
oxocene 139 was previously transformed into the alcohol 140 in an excellent overall yield and stereoselectivity. With this compound in hand,
treatment with t-BuOCl or NBS efficiently furnished the corresponding chloro and bromo bicyclic ethers,
which were prepared for the key CM reaction by transformation into the terminal olefins
141 and 142, respectively. For the E-selective installation of the enyne moiety, they carried out a CM reaction of compounds
141 and 142 with crotonoaldehyde in the presence of the Grubbs II catalyst (4) to yield the corresponding (E)-α,β-unsaturated aldehydes, which after a final reaction with lithium TMS-diazomethane
delivered the final products 135 and 137 in 74% and 61% overall yields, respectively. For the synthesis of the Z-isomers, they applied the Lee methodology,[99] based on a cross metathesis with enyne 145 in the presence of the H-G II catalyst (6) for the direct delivery of the Z-isomers of both natural products 134 and 136 in good yields, albeit in modest stereoselectivities in favor of the Z-isomer (3.8:1 mixture for 134; 4:1 for 136) (Scheme [12]). Finally, the described stereodivergent strategy to the chlorofucins and bromofucins
allowed confirmation of the absolute configurations of these natural products by comparison
of the spectroscopic properties and optical rotations of the synthetic compounds with
those reported for the natural products.
Scheme 12 Total synthesis of chlorofucins and bromofucins
Metathesis-based methodologies have been used in combination with other innovative
technologies in organic synthesis laboratories. Indeed such a case is continuous flow
technology, which has been recently implemented in combination with metathesis reactions
to enhance their synthetic value.[100]
[101] An interesting application of this can be found in the total synthesis of neomarchantin
A (146), a natural product belonging to the macrocyclic bisbibenzyl family, which has been
paid much attention from chemical and biological standpoints due to its intriguing
structural and biological features. For example, these compounds exhibit a wide range
of biological activities, including antibacterial, antimycotic, antitumoral, and antiviral
activities. The total synthesis of neomarchantin A (146) (Scheme [13]), described by Collins and co-workers,[102] represents the first synthesis of this natural product, which is based on a RCM
reaction of the diolefin 147, to obtain the corresponding macrocyclic derivative in a modest 43% yield when H-G
II catalyst (6) was employed in toluene at 110 °C. Interestingly, when the cyclization reaction
was done in continuous flow, the yield of the macrocyclic product could be increased
to a 49% yield. Final hydrogenation and methyl ether cleavage afforded the targeted
natural product in 88% overall yield (Scheme [13]).
Scheme 13 Total synthesis of neomarchantin A
As a final example for this section, an interesting strategy was delineated for the
synthesis of the cyclopropane fatty acid (+)-majusculoic acid (148), an interesting secondary metabolite with significant antifungal activity, by Zhang
and co-workers (Scheme [14]).[103] Accordingly, they devised a stereoselective cyclopropanation from the C
2-symmetric 14-membered dilactone according to previously performed DFT calculations
that revealed a preference of the β-faces of both olefins for the attack of the corresponding
cyclopropanating reagents. For the synthesis of the required dilactone 150, they envisioned an unusual RCM dimerization from diolefinic monomer 149. To this aim, after an extensive study in which an array of catalysts, solvents,
and temperatures were screened, they found the Grubbs II catalyst (4) in refluxing dichloromethane as the best reaction conditions to obtain the dilactone
150 (53%), accompanied by the monomeric RCM product 151 in 26% yield. With the dilactone 150 in hand, it was possible to demonstrate the reliability of the theoretical predictions
regarding the stereochemical outcome of the cyclopropanation. Indeed, the formation
of the biscyclopropane 152 was realized in 80% yield as a single stereoisomer when 150 was treated with diazomethane in the presence of palladium(II) acetate. The dimerization
process was then continued by a dedimerization process by opening of the dilactone
152 with N,O-dimethylhydroxylamine and reduction of the resulting Weinreb amide to the aldehyde
153. In a final sequence of eight steps, in which the awkward (E,Z)-bromodiene system was installed via their newly developed new Kocienski–Julia-type
reagent 154, the synthesis was complete, obtaining a product which matched all the spectroscopic
and physical properties with those reported for the natural product, except for the
optical rotation, which was the opposite. Therefore, this asymmetric synthesis allowed
the confirmation of the absolute configuration for the natural product that corresponded
to ent-148, as initially proposed (Scheme [14]).
Scheme 14 Total synthesis of (+)-majusculoic acid
2.2
Total Syntheses Based on a Cross-Metathesis Reaction
In contrast to the RCM reactions, the cross-metathesis (CM) reaction has received
less attention in the field of total synthesis, probably due to the inherent difficulties
of the intermolecular assembly of two alkenes, in which dimerization processes compete
substantially. Despite these difficulties, the appealing features of the metathesis
reactions, in terms of functional group tolerance and mild reaction conditions, have
prompted intense research activity with the goal of minimizing the chemo- and stereoselective
issues associated with this modality. As a consequence, CM reactions have emerged
as a common chain-elongation method to append small side chains or to assemble linear
complex molecular frameworks. Concerning the stereoselectivity of the reaction, the
thermodynamic control that governs these reactions leads to the preferred formation
of the E-alkene. Within the field of total synthesis, several interesting syntheses of natural
products have been reported in which the CM represented the key step for the coupling
of the important fragments of the molecule, thus allowing the completion of the synthesis.
Among the most outstanding contributions of the last few years are the total syntheses
of the fluorinated cryptophycin (155),[104] FD-891 (156),[105] murisolin (157),[106] ledoglucomides A (158) and B (159),[107] epicoccamide D (160),[108] synargentolide B (161),[109] homaline (162),[110] bitungolide F (163),[111] penarolide (164),[112] cytospolide D (165),[113] goniocin (166),[114] cryptomoscatone F1 (167),[115] curvicollide C (168),[116] proposed structures for cryptorigidifoliol K 169a–d,[117] and alotaketal A (170)[118] (Figure [6]).
Figure 6 selected natural products synthesized via a CM reaction (yields refer only to the
CM reaction)
However, there are some specific cases that deserve a more detailed description. As
a first example within this category is the formal total synthesis of spirangien A
(171), a highly potent cytotoxic spiroketal isolated from the prolific myxobacterium Sorangium cellulosum. Among the various total syntheses described for this natural product, that reported
by Rizzacasa and co-workers (Scheme [15])[119] utilized a CM reaction in the union of the olefins 172 and 173 under the assistance of the Grela–Grubbs–Hoveyda catalyst (10). This CM reaction afforded the E-enone 174 in an excellent 80% yield, providing a straightforward alternative to the use of
a Horner–Wadsworth–Emmons reaction, which requires additional steps for the preparation
of the required phosphonate. In this way, the backbone of the spiroketal core found
in the natural product was constructed, requiring only six additional steps for the
rapid and efficient preparation of the spiroketal precursor 176. From the advanced spiroketal precursor 176, the completion of the natural product was reported earlier by Paterson and co-workers
(Scheme [15]).[120]
Scheme 15 Formal total synthesis of spirangien A
Another case of note is represented by the total synthesis of mueggelone (177), an inhibitor of fish embryo larval development. After several total syntheses reported
for this interesting natural product, some of which required around 20 steps, Meshram
and Kumar[121] described a convergent total synthesis utilizing only eight steps in the longest
linear sequence, featuring a final CM step in which olefins 178 and 179 were assembled by treatment with the Grubbs II catalyst (4) to directly deliver the natural product 177 in 40% yield (Scheme [16]).
Scheme 16 Total synthesis of mueggelone
The prominent antitumor activities of spliceostatin E (180) and related compounds, such as spliceostatin A (181) and FR901464 (182), have elicited great interest in their syntheses. Among them, the total syntheses
developed by Ghosh and co-workers is based on a CM reaction,[122] representing a brilliant example of the power and robustness of the methodology
(Scheme [17]).
Scheme 17 Total synthesis of spliceostatins E and A and FR901464
The ability of these natural products to inhibit the cellular splicing process, an
essential step for the gene expression, renders them a novel class of anticancer products.
In fact, spliceostatin E (180) and FR901464 (182) exhibit potent anticancer activities in the ranges from 1.5 to 4.1 and from 0.6
to 3.4 nM, respectively, against various human cancer cell lines. The syntheses of
both compounds were achieved via a final key CM that efficiently joined fragments
183 and 184 for spliceostatin E (180)[122] and 185 and 186 for spliceostatin A (181),[123] in 71% and 57% yields, respectively, with both reactions displaying exquisite chemo-
and stereoselectivities. A final acidic treatment of compound 181 afforded the natural product 182 (Scheme [17]).
Scheme 18 Total synthesis of mycalolides A and B
An excellent example that illustrates the advantage that the CM reaction can offer,
especially for macrocyclic compounds, over a RCM reaction, are the synthesis of the
natural products mycalolide A (187) and B (188) by Kigoshi, Kita, and co-workers (Scheme [18]).[124] Both compounds, isolated from the marine sponge Mycale sp., are members of a family of natural products characterized by the presence of
a trisoxazole macrocycle with potent cytotoxic activities as a consequence of actin-depolymerizing
activity. A first attempt at the total synthesis was carried out by them via a RCM
reaction of the diolefins 189a and 189b. However, despite this reaction affording the macrocyclic derivatives 190a and 190b in reasonably good yields (40–76% depending on the employed catalysts and solvents),
the reaction was not stereoselective in either case, providing a mixture of the E/Z isomers in ratios of 1:1 to 1.9:1.0. This proportion could be increased to a 2.7:1.0
ratio in favor of the required E-olefin when the desilylated derivative 189b was employed as the starting precursor. As an alternative, they attempted a CM/macrolactonization
approach which proved to be more efficient. Thus, when a mixture of olefins 191 and 192 was subjected to the catalytic action of the H-G II catalyst (6) in refluxing dichloromethane, the olefin 193 was obtained in an impressive 77% yield and in an improved 5:1 mixture of the E/Z isomers. After the successful CM step, a Yamaguchi macrolactonization was conducted
to prepare the corresponding macrolactone in a remarkable 77% yield, which was finally
directed to the natural products mycalolides A (187) and B (188) in ten additional steps (Scheme [18]).
Scheme 19 Total synthesis of tiacumicin B aglycone
The scope of the CM reaction has been expanded to conjugated dienes by Altmann and
Glaus in 2015[125] during their synthetic studies directed towards the aglycone of tiacumicin B (194), a macrocylic antibiotic with potential application against Mycobacterium tuberculosis and whose total synthesis was not reported when the studies were published (Scheme
[19]). Thus, combination of olefin 195 and the conjugated diene 196 in the presence of the H-G II catalyst (6) provided a 56% yield of compound 197 as an inseparable mixture of E/Z isomers in a 6.7:1 ratio. The preparation of ester 199 was followed by a Suzuki macrocyclization, to afford the macrocyclic aglycone derivative
of tiacumicin B (200) in 73% yield. Global desilylation of the macrocyclic derivative 200 was performed with Et3N·3HF to obtain tiacumicin B aglycone (194). In 2015, Gademann and co-workers reported the total synthesis of tiacumicin B,
also named fidaxomicin or lipiarmycin A3 (53), based on a RCM strategy, as depicted in Figure 3.[59] In addition to the synthetic studies by Altmann and Glaus,[125] closely related strategies have been reported by Zhu and co-workers (2015)[126] and Roulland and co-workers (2017),[127] also using CM reactions (Scheme [19]). The former achieved the assembly of olefin 201 and diene 202, while the latter carried out the assembly of olefin 204 with diene 205 to yield the corresponding CM products 203 and 206 in 38% and 39% yields, respectively and with stereoselectivities comparable to those
obtained by Altmann and Glaus. However, unlike the studies of Altmann and Glaus, in
these cases they found serious difficulties, mainly with the subsequent removal of
the employed protecting groups within the products 203 and 206, thus forcing them to explore different synthetic strategies in both cases, to overcome
these synthetic hurdles and to reach the targeted tiacumicin B aglycone (194).
Another relevant contribution of the use of the CM reaction in total synthesis was
published in 2016 by Takahashi and co-workers for the total synthesis of aromin (207) (Scheme [20]),[128a] a member belonging to the wide family of the Annonaceous acetogenins , isolated from the Annonaceae plants and which exhibits a broad spectrum of biological activities including anticancer,
antibiotic, immunosuppressive, antifeedant, and pesticidal activities. Aromin (207), isolated together with aromicin in 1996 by McLaughlin and co-workers,[129] has been recognized as an antitumoral compound with a significant cytotoxicity profile
against various tumor cell lines, albeit lower compared to other acetogenins. The
synthetic work by Takahashi and co-workers,[128a] based on their previous synthetic work,[128b] is interesting for two reasons. On one hand, they demonstrated the utility of the
CM reaction by using the Grubbs II catalyst (4) to connect a terminal olefin and an enone to construct the corresponding linear
E-enone. On the other hand, this work also led to the structural revision of aromin
(207). Thus, the assembly of olefin 208 and enone 209 proceeded smoothly when they were exposed to the Grubbs II catalyst (4) to provide the E-enone 210 in 68% yield and excellent stereoselectivity. With this compound 210 in hand, they accomplished the completion of the proposed structure for aromin in
four additional steps in 73% overall yield. Surprisingly, the spectroscopic and physical
properties of the synthetic material 207 were notably different from those reported for the natural aromin. According to the
detected discrepancies found in the 1H NMR spectra, they suggested a structural difference around the central THF ring
and, for this reason, they prepared the stereoisomer 211, according to a similar synthetic sequence as for 207, but once again, the NMR data were inconsistent with those of the natural product.
After an exhaustive comparative analysis of the NMR data of a wide set of acetogenins
possessing a 4-hydroxy group adjacent to the γ-lactone ring, together with a detailed
analysis of the MS fragmentations, they were led to propose the structure of the natural
product montanacin D (212), whose synthesis was similarly accomplished by them according to the same CM strategy,
for the revised structure for aromin (Scheme [20]).[128]
Scheme 20 Total synthesis of proposed structure for aromin and its structural revision
Another interesting application of the CM reaction was the recent total synthesis
of depudecin (213) by Sarabia and co-workers,[130] in which functional group tolerance provided by the employed catalysts was verified,
in particular with the especially sensitive oxirane rings (Scheme [21]). Depudecin represents an unprecedented natural product that selectively inhibits
histone deacetylases I and II in the low μM range. With respect to its total synthesis,
the construction of the complete framework of depudecin was accomplished via an efficient
CM reaction of epoxy olefin 214 and diepoxy olefin 215 by using the H-G II catalyst (6) to obtain the resulting triepoxy olefin 216 in a reasonable 50% yield, exclusively as the E-isomer. With the product 216 in hand, its transformation into the final natural product was successfully achieved,
requiring the reductive opening of the terminal epoxy alcohol, carried out in three
steps, and final removal of the protecting group of the resulting product 217 (Scheme [21]). It is important to point out that this total synthesis represents the third total
synthesis reported thus far for this interesting molecule, resulting in an improvement
upon the previous syntheses reported by Schreiber and co-workers in 1990,[131] and also by Sarabia and co-workers group utilizing a linear strategy.[132] In addition, the new convergent CM synthetic strategy provided rapid access to different
stereoisomers of depudecin, which allowed for an evaluation of the influence of the
stereochemistry upon biological activities.
Scheme 21 Total synthesis of depudecin
A final example within this section is represented by the interesting synthetic strategy
reported by Pietruszka and co-workers for the synthesis of putaminoxins B/D (218), identified as a C-9 epimeric mixture and whose absolute configurations were not
yet established (Scheme [22]).[133] In order to unambiguously establish their absolute configurations, the four possible
stereoisomers 218a–d were prepared by a CM reaction of dimers (S,S)- and (R,R)-219, previously prepared by a dimerization CM reaction of the corresponding monomers,
with olefins (S)- and (R)-220. The choice to use the dimers 219 instead of the corresponding monomeric olefins was justified because better yields
were obtained for the desired cross-metathesis products 221 versus CM reactions with the monomers, in which the dimeric products 219 were the main compounds detected. Indeed, when each isomer of 219 was treated with the corresponding (S)- and (R)-220 in the presence of the Grubbs II catalyst (4) the corresponding products 221a–d were obtained in 36%, 72%, 24%, and 83% yields, respectively. Then, each isomer was
subjected to a sequence that included basic treatment, a Yamaguchi macrolactonization
and final THP protecting group removal to provide all the possible stereoisomers of
the macrolactone 218 in high yields. With all the isomers in hand, an extensive spectroscopic study, in
which they compared the NMR data collected from all these compounds with those reported
for the natural products, resulted in disappointment. This was due to the inability
to correlate the resulting NMR data of the synthetic compounds with those reported
for the natural products, thus leading them to reconsider the initially proposed structures
for putaminoxins B/D. Intriguingly, after a thorough inspection of the NMR data of
related nonenolides, they found that the data reported for the natural products hypocreolide
A (222) and aspinolide A (223) matched with those reported for the natural products putaminoxins B/D (Scheme [22]).
Scheme 22 Total synthesis and structural studies of putaminoxins B/D
2.3
Strategies for Selective and Efficient Metathesis Reactions of Alkenes
2.3.1
Temporary Tethered Ring-Closing Metathesis
A highly inventive and clever strategy to transform the intermolecular character of
the cross-metathesis of alkenes into an intramolecular variant has been explored by
means of the development of temporary tethers, through which the corresponding olefinic
partners can be transiently coupled. This strategy not only provides the advantage
of improved efficiency, which is a feature of RCM versus the CM reaction, but also
the possibility of controlling the geometry of the resulting double bond to favor
the Z-olefin when a strained ring is temporarily formed. However, this strategy is limited
to molecular structures in which a possible linkage of the tether system is present,
usually in form of hydroxy groups at allylic or homoallylic positions. Due to the
versatility, stability, and ease of cleavage, silicon-based tethers have been the
most frequently employed for this strategy. In fact, the temporary silicon-tethered
RCM (TST-RCM) was described initially by Grubbs and Fu in 1992.[134] Thus, in order to favor the formation of the Z-isomer, a temporary [7]-membered ring is required, for which the silicon tethered
is represented by precursor A. In case of precursor B, whose RCM would deliver an [8]-membered ring, the preference would be the formation
of the E-isomer (Scheme [23], A). In 2012, the applications of silicon-tethered RCM in synthesis of natural products
was reviewed[135] and, as a consequence, in this section, we will focus on recent examples published
after 2012, together with other cases of tethered RCM reactions based on other tethers
and their applications in the total synthesis of natural products. Some relevant contributions
to this field are those by Kobayashi,[136] Hoye,[137] and Evans,[138] as well as numerous syntheses of natural products belonging to the family of 6-substituted
5,6-dihydro-α-pyrones, such as hyptolide (224),[139] pectinolide C (225),[140] and umuravumbolide (226),[141] all of which have been efficiently achieved according to this strategy from precursors
227, 228, and 229, respectively (Scheme [23], B).
Scheme 23 General temporary silicon-tethered RCM (TST-RCM) reactions and selected natural products
synthesized via this strategy
In addition to silicon-tethered RCM reactions containing an O–Si–O linkage, this strategy
has also been utilized with substrates containing an O–Si–C linkage, in which an oxidative
or base-induced ring cleavage step is required after the RCM reaction. Prominent amongst
these cases is the 2013 synthesis of the potent cytotoxic natural product amphidinolide
V (230) by Lee and Volchkov (Scheme [24]),[142] in which the stereocontrolled construction of the existing double bonds was devised
utilizing a tethered RCM of silyl ether 234 that provided the RCM product 235 in an excellent 96% yield, which was then subjected to an allylic 1,3-transposition
mediated by Re2O7 to afford siloxane 236 as an 85:15 E/Z inseparable mixture in 85% yield. The presence of the trimethylsilyl group in the
1,3-diene subunit was proven to be critical to prevent unwanted side reactions during
the metathesis reaction. The RCM precursor 234 was prepared beforehand via an enyne RCM of silyl ether 231 to furnish a cyclic siloxene type A, according to the general reaction depicted in Scheme [24], which was opened by treatment with methyllithium, to deliver compound type B. The resulting alcohol 232 was then coupled with silane 233 via a dehydrogenative reaction, mediated by a catalytic amount of (Xantphos)CuCl
and in the presence of lithium tert-butoxide, to afford the targeted precursor 234 in an excellent yield and exquisite control of the geometry of the double bonds.
Promptly, with compound 236 in hand, they proceeded with the connection of the other important fragment of the
molecule via a cross-aldol condensation to provide aldehyde 237 in a sequence of seven steps, followed by a Yamaguchi macrolactonization that provided
macrocyclic derivative 238. The completion of the synthesis from this advanced precursor was then executed in
five steps, consisting of a Sharpless asymmetric epoxidation from the corresponding
allylic alcohol, reductive opening of the resulting diepoxy alcohol, and final removal
of the trimethylsilyl group.
Scheme 24 Total synthesis of amphidinolide V
The use of this strategy for the selective construction of a trisubstituted Z-olefin is nicely illustrated in the synthesis of the important antimitotic agent
epothilone B (239), in which the required Z-olefin at the C12–C13 bond has been a critical issue, and which different solutions
across the vast number of reported synthesis of this natural product have been found.[143] Thus, whereas a conventional RCM of an advanced diolefin precursor was devoid of
stereoselectivity, a silicon-tethered RCM reaction initially explored by Mulzer and
Gaich provided the required Z-isomer with a remarkable 5:1 selectivity albeit through a long synthetic sequence.[144] In a shorter and rapid approach, Lin and co-workers explored the RCM reaction through
the formation of a bissiloxane-tethered precursor, to secure the double bond geometry
(Scheme [25]).[145] To this aim, the precursor 242, efficiently prepared by joining segments 240 and 241, was subjected to the Grubbs II catalyst (4) to yield the corresponding RCM product 243 in nearly quantitative yield (95%) as a 1.7:1 mixture of Z/E isomers, which could be separated after selective cleavage of the silicon-tethered
by reaction with HF·pyridine. Notably, despite the moderate selectivity of the RCM
reaction, the required aldehyde 244 in the form of the pure Z-isomer was obtained in 40% over three steps, representing the most step-economic
synthesis reported thus far. From aldehyde 244, the completion of the synthesis of epothilone B was accomplished in an efficient
manner in six steps, with a TiCl4-mediated aldol reaction and a Yamaguchi macrolactonization as the key steps (Scheme
[25]).
Scheme 25 Total synthesis of epothilone B
Scheme 26 Total synthesis of amphidinolides T
The extension of the concept of a temporary tethered RCM to other groups such as esters
or acetals would expand the synthetic opportunities that this strategy might offer.
One such application is elegantly illustrated with the total syntheses of amphidinolides
T1, T3, and T4 by Clark and Romiti (Scheme [26]).[146] As for amphidinolide V, discussed above, the amphidinolides T1, T3, and T4 are also
macrolides isolated from marine dinoflagellates with potent cytotoxic activities.
Structurally, the so-named amphidinolides of the T series are characterized by the
presence of a trisubstituted tetrahydrofuran. In the case of amphidinolide T1 (245), the synthesis of the C1–C11 fragment was initially attempted by utilization of
a CM reaction to connect the corresponding olefins, but all attempts were thwarted
to obtain the desired CM product. The implementation of a tethered RCM strategy to
prepare this fragment was then attempted via a temporary ester, for which diolefin
ester 246 was prepared, but again, the reaction was found to be unsuccessful, instead obtaining
the RCM product 247. In light of these discouraging results, the use of a salicylate spacer as an alternative
ester-tether RCM was explored; indeed the treatment of 248 with the Hoveyda–Grubbs I catalyst (5) afforded an isomeric mixture (E/Z 1.2:1) of the lactone 249 in an impressive 96% yield. The removal of the spacer group was carried out by base
treatment, followed by a chemoselective hydrogenation to complete the western fragment
of the natural product in form of the product 250. The coupling of this fragment with the eastern fragment provided access to the corresponding
seco acid, which was subjected to a Yamaguchi macrolactonization to provide 251, representing the macrocyclic core of the amphidinolides T. The completion of the
synthesis of this natural product was achieved with the hydrosilylation of the alkyne
present in 251 using the ruthenium catalyst [Cp*Ru(MeCN)3]PF6 that afforded a 1:1 mixture of isomeric (Z)-vinylsilanes 252 and 253 that could be separated by chromatography. Whereas, 252 was transformed into amphidinolide T1 (245) via epoxidation of the vinylsilane, followed by a Fleming–Tamao oxidation, the product
253 provided amphidinolides T3 and T4 under the same synthetic sequence (Scheme [26]).
The ester-tethered RCM was similarly used in the synthesis of polyacetylene-type metabolites,
such as the atractylodemaynes C (254) and F (255) by Schmidt and Audörsch (Scheme [27]),[147] as a way of controlling the geometry of the depicted double bonds. According to
a tandem RCM/base-induced eliminative ring opening, carried out in one pot, they were
able to construct a E,Z-diene derivative, found in these natural products. Thus, when compound 256 was treated with the Grubbs II catalyst (4), followed by NaHMDS and Meerwein’s salt, compound 257 was obtained in 73% overall yield through intermediates A, B, and C. Starting from the resulting ester 257, the installation of the enediyne moiety of the atractylodemaynes C and F was achieved
without issues via alkynyl homologation and a Sonogashira coupling (Scheme [27]).
Scheme 27 Total synthesis of S-atractylodemayne C and F
As an alternative to the silicon-tethered RCM reaction, the use of acetals to connect
both olefinic partners has also been developed and applied in 2013 in the synthesis
of a stereoisomer of squamocin K (258), the 14,21-di-epi isomer 259 by Hou and co-workers (Scheme [28]).[148] Squamocin K, as other Annonaceous acetogenins, displays a wide range of biological activities, such as antitumor, antiparasitic,
pesticidal, antimicrobial, and immunosuppressive activities, by virtue of its inhibition
of mitochondrial complex I. In this synthetic proposal, precursor 260, readily prepared from a C2 symmetric diene diol, was treated with the Grubbs II
catalyst (4) to obtain the resulting RCM product, as a result of a multiple RCM process. In a
second step, the assembly of the resulting polyene and 10-chlorodec-1-ene was induced
by treatment again with the Grubbs II catalyst (4) to deliver the complete framework of this class of natural products in form of compound
261. Finally, the completion of the synthesis of this stereoisomer of squamocin K was
achieved in 12 additional steps, including the formation of the required tetrahydrofuran
rings, via previous activation of the resulting hydroxy groups as mesyl derivatives
and subsequent intramolecular displacements to afford 262, and the final introduction of the unsaturated lactone (Scheme [28]).
Scheme 28 Total synthesis of 14,21-di-epi-squamocin K
A unique tether for a RCM reaction was employed by Hanson and co-workers in the synthesis
of strictifolione (263) (Scheme [29], A),[149] a natural product isolated from the stem bark of Cryptocaria stritifolia with antifungal activities. In their synthesis, they used a phosphate-tether that
allowed in one pot, a sequential RCM, CM, and a chemoselective hydrogenation process,
followed by a one-pot, sequential reductive allylic transposition/tether removal and
final CM. To this end, olefinic phosphate 264 was treated with the H-G II catalyst (6) and then, after solvent evaporation, addition of cis-stilbene in DCE and heating at 60 °C for two hours. The subsequent chemoselective
diimide reduction provided phosphate 265 in 52% overall yield. The following one-pot protocol was initiated with the reaction
of 265 with Pd(OAc)2 in the presence of formic acid, cesium carbonate, and PPh3 to produce a reductive allylic transposition, followed by addition of methyl sulfate
and treatment with LiAlH4 to furnish diol 266 in 65% overall yield. A final CM reaction with vinyl lactone 267 in the presence of the H-G II catalyst (6) provided the natural product strictifolione (263) in 77% yield (Scheme [29], A). In a similar strategy, they described the synthesis of the antifungal macrolide
Sch-725674 (49) (Scheme [29], B),[150] wherein the phosphate 268 was treated in a similar one-pot RCM/CM/chemoselective hydrogenation sequence as
described for 264, with the use of alkene 269 to obtain compound 270 in 59% yield. The removal of the phosphate tether was accomplished by treatment with
LiAlH4 and, then, after a sequence of transformations, including selective protection of
the 1,3-diol system as an acetal, Sharpless asymmetric epoxidation of the resulting
allylic alcohol, selective tosylation of the primary alcohol, and introduction of
the acryloyl unit, provided compound 271. This compound was set up for a reductive opening process, which was carried out
in one pot by sequential treatment with NaI, Zn, and acidic work-up, to deliver diolefin
272. In contrast to a direct RCM reaction of diolefinic triol 272, carried out by Prasad and co-workers (see Figure 3),[55a] that afforded the natural product in a modest 36%, Hanson and co-workers found that
prior protection of compound 272 as the MOM derivative resulted in a more efficient RCM process to provide the final
product 49 in 84% yield after removal of the protecting groups with TFA (Scheme [29], B).
Scheme 29 Total synthesis of strictifolione and Sch-725674
2.3.2
Relay Ring-Closing Metathesis
Since the pioneering work of the Hoye group on the relay ring-closing metathesis (RRCM)
concept in 2004,[151] and the first application in total synthesis by the Porco group in the same year,[152] many syntheses have benefited enormously from this strategy as a method to surmount
the instances in which RCM reactions proved ineffective or sluggish, mainly due to
steric factors, as well as electronic factors.[153] A representative example to illustrate the implementation of this strategy and its
dramatic effect can be found in the 2012 synthesis of the cytotoxic natural product
penostatin B (273) by the Shishido and co-workers (Scheme [30]).[154] Having devised a RCM strategy for the preparation of the dihydropyran system contained
in the natural product, the examination of this reaction with the acyclic precursor
274 by using various catalysts and different reaction conditions provided the desired
unsaturated δ-lactone 275 in a modest 46% yield, as the best case, when the H-G I catalyst 5 was employed. In an effort to improve upon this yield, they made use of the relay
RCM strategy, for which precursors 276 and 277 were prepared. The treatment of these compounds with the H-G I catalyst 5 provided the RCM product 278 in improved 78% and 83% yields from 276 and 277, respectively, demonstrating the synthetic value of this strategy. With compound
278 in hand, the introduction of the alkenyl appendage was successfully achieved from
acetyl pyranoside 279, via vinylstannane 280 in a highly stereoselective manner. Finally, the manipulation towards the final product
was undertaken in six additional steps.
Scheme 30 Total synthesis of penostatin B
In the development of an enabled and flexible strategy for the synthesis of the cyclodepsipeptidic-like
natural products related to the jasplakinolide/geodiamolide family, Arndt, Waldmann,
and co-workers designed a divergent solid phase synthesis for this class of compounds
based on a unique RRCM strategy on solid phase to construct the macrocyclic core,
directing the catalyst’s action to the required break point (Scheme [31]).[155] With the preparation of a PS resin for the linkage of the peptidic chain, which
carried a diene unit, the acyclic precursor loaded onto the resin (resin 281) was synthesized in an efficient manner. The subsequent reaction with the Grubbs
II catalyst (4) delivered the protected cyclodepsipeptide 282 in 34% yield as a 1:1.2 mixture of a separable E/Z isomers, through the formation of the carbene intermediate A with the concomitant release of resin 283. The removal of the protecting groups of each pure isomer afforded natural product
seragamide A (284) and its Z-isomer. In a similar fashion, a collection of jaspamide analogues were generated
for biological studies as antitumor agents.[155]
Scheme 31 Total synthesis of seragamide A and related cyclodepsipeptides
2.3.3
Stereoselective Alkene Metathesis
The reversible nature of olefin metathesis represents a practically insurmountable
barrier to access to the often energetically less favored Z-olefin. As a consequence, the design and development of catalysts capable of providing
high Z selectivity, either in RCM as in CM reactions, represents an important challenge.
To this aim, initial efforts by Schrock, Hoveyda, and co-workers have been focused
largely on the modification of the ancillary ligands bound to the metal center and
have led to the identification of the first Z-selective catalysts based on molybdenum and tungsten (catalysts 15–22 in Figure [1]).[156] In this family of catalysts, the introduction of a bulky aryloxy moiety forces the
substituents on the generated metallacyclobutane intermediate A (Scheme [32]) all syn to yield the Z-olefin in a kinetically controlled process. On the other hand, the advent of ruthenium-based
catalysts (e.g., catalyst 23 in Figure [1]), in which bulky aryl groups are introduced on the N-heterocyclic carbene, by Grubbs
and co-workers, allowed similar access to a Z-selective process through intermediates type B.[157] Despite the formation of the E-isomers being generically favored in these reactions for thermodynamic reasons, there
are numerous cases in which the small energy difference between the E and the Z isomers results in a mixture of both geometric isomers. In view of this situation,
the development of catalysts that promote kinetically E-selective processes represents a new challenge. In response to this requirement,
Grubbs and co-workers have described the first catalysts capable of generating E-olefins starting from E-olefins as the reactants.[158] These catalysts are a new generation of catalysts that are termed stereoretentive
olefin metathesis catalysts, and they feature the presence of a cyclic catecholthiolate
unit (catalysts 24–27 in Figure [1])[22]
[159] and proceed through intermediates of type C. To this arsenal of valuable and useful catalysts developed in the last few years
one must also add chiral catalysts, with which it is possible to perform asymmetric
olefin metathesis,[61,160] as described in Scheme [8]. A detailed description of the design and development of these new selective catalysts,
as well as their applications in the synthesis of natural products, has been widely
covered in various reviews.[161] Nevertheless, we would like to summarize in this section relevant examples of selective
RCM and CM reactions as representative applications that prove the synthetic validity
and potential of this new generation of catalysts. Thus, Scheme [32] summarizes examples of Z-selective RCM reactions[156]
[162] and Scheme [33] shows cases of Z-selective CM reactions[163] in the field of the total synthesis.
Scheme 32 Applications of Z-selective RCM reactions in total synthesis
Particularly relevant is the synthesis of prostaglandin F2α (303) according to the 2017 strategy developed by Hoveyda and co-workers (Scheme [33])[163f] based on the original and brilliant concept of methylene capping. This strategy
was designed to broaden the scope of the Z-selective cross-metathesis protocols, which are inefficient with olefins containing
allylic or homoallylic alcohols, aryl groups, aldehydes, or carboxylic acid substituents.
Having identified (Z)-but-2-ene (302) as a suitable capping agent, respective treatments of trihydroxy olefin 300 and unsaturated carboxylic acid 301 with alkene 302 and the Ru dithiolate catalyst 25, followed by mixing and treatment with additional catalyst 25 under reduced pressure, afforded prostaglandin F2α (303) in 59% yield and >98:2 selectivity in favor of the desired Z-olefin. The use of this capping agent avoids the formation of the unstable methylidene
species in favor of a more stable substituted carbene, allowing the use of substrates
without resorting to protecting groups. This elegant strategy was similarly used for
the synthesis of Z-macrocyclic alkenes.
Scheme 33 Applications of Z-selective CM reactions in total synthesis
An outstanding application of these selective catalysts in the synthesis of complex
natural products is the synthesis of disorazole C1 (305), a secondary metabolite that displays excellent anticancer and antifungal profiles,
described by Hoveyda and co-workers (Scheme [34]).[164] The presence of the (Z,Z,E)-1,3,5-triene unit demands an exquisite level of geometric control for the stereoselective
synthesis of this fragment. In an initial approach, they prepared the Z-vinyl iodide 307 and the E,Z-diene 309 in excellent yields and complete stereoselectivity by using the catalyst 16 from the terminal olefins 306 and 308, respectively. However, after coupling of both fragments via a Suzuki reaction, the
subsequent dimerization process failed to form the resulting coupling product. Therefore,
they decided to assemble the acid derived from ester 308 and vinyl iodide 310 and the resulting ester 311 was transformed into the Z-boronic ester 312, mediated by the catalyst 18. With this compound in hand, they conducted the dimerization process in a carefully
optimized reaction, where the choice of palladium catalyst, base, and solvent was
critical to obtaining a good yield (60%) of the resulting [30]-membered ring protected
disorazole C1. Final desilylation provided the coveted natural product 305.
Scheme 34 Total synthesis of disorazole C1
2.3.4
Alkene Metathesis in Tandem Reactions
The design of suitable multifunctional molecules that enable the triggering of a cascade
of events, including a sequential alkene metathesis process or a combination of an
alkene metathesis with other types of reactions in a well-defined order would provide
a formidable increase in structural complexity in a single operation. As a consequence,
we can find in the literature a large number of total syntheses and synthetic approaches
in which alkene metathesis processes in cascade sequences, by combination of ROM/RCM
or RCM/CM reactions, have been applied.[165] An initial interesting synthetic application of these processes in total synthesis
is found in the formal total syntheses of dysiherbaine (313) and neodysiherbaine A (314) by Lee and co-workers (Scheme [35])[166] through an elegant ROM/RCM tandem process from bicyclic compound 315, which was prepared in enantiomerically pure form through a stereoselective Diels–Alder
reaction, followed by resolution of the racemic mixture. This compound 315 underwent an initial ROM process that delivered a ruthenium intermediate A, followed by a RCM reaction when subjected to the Grubbs II catalyst (4) to yield bicyclic derivative 316, which represents the core skeleton of dysiherbaine and related compounds with the
correct relative stereochemistry. The need to selectively oxidize the cyclic olefin
in the presence of the terminal alkene found in compound 316 forced them to increase the difference of reactivities between the olefins by the
preparation of the enol acetate 317, prepared in the same way as 316, but in the presence of vinyl acetate and using H-G II catalyst (6) for a final CM with vinyl acetate. In this way, compound 317 was obtained in an excellent 95% yield, generating a very well differentiated electronic
environment between both alkenes for further selective functionalization. In fact,
they succeeded in the preparation of advanced precursor 318 via transformation of the enol acetate into a temporary alcohol, followed by dihydroxylation
of the cyclic olefin, protection of the resulting diol as an acetal, oxidation of
the primary alcohol to the acid, and esterification. The resulting product 318 was utilized by Sasaki and co-workers (Scheme [35])[167] in their total syntheses of the natural products dysiherbaine (313) and neodysiherbaine A (314), thus representing the formal synthesis of both natural products.
Scheme 35 Formal total synthesis of dysiherbaine and neodysiherbaine A
Scheme 36 Total synthesis of clusianone
Particularly interesting are the syntheses of clusianone (319) and clavilactone A (323) which utilize combinations of cascade processes of metathesis reactions. In the
case of clusianone (319), a natural product isolated from C. congestiflora with antiviral activity against HIV and Epstein-Barr virus, its synthesis was envisaged
by Plietker and co-workers (Scheme [36])[168] from tricyclic derivative 322, prepared from bicyclic compound 321 via an allylation/intramolecular Claisen condensation/benzoylation sequence. Thus,
when 322 was treated with the Grubbs II catalyst (4) in the presence of 2-methylbut-2-ene, the natural product clusianone (319) was obtained in 65% yield as the result of a tandem ROM/CM process. The preparation
of the bicyclo[4.3.1]decenone derivative 321 via a RCM reaction from 320, was conceived by them as a way of controlling the stereochemical outcome in favor
of the desired cis-relative configuration of the final product due to the conformationally restricted
environment imposed by the bicyclic system. The subsequent ROM reaction, followed
by a CM reaction, in the presence of 2-methylbut-2-ene, allowed the unmasking of the
prenylated side chains present in the final product (Scheme [36]). The same strategy was employed by them in the synthesis of guttiferone A.[169] For the clavilactones, a family of natural products structurally characterized by
a rigid 10-membered macrocycle fused to a hydroquinone and to an α,β-epoxy-γ-lactone,
the cyclobutenecarboxylate 324 was devised by Takao and co-workers as an appropriate precursor to promote a tandem
ROM/RCM reaction mediated by a ruthenium-based catalyst (Scheme [37]).[170] In fact, when 324 was treated with the Grubbs I catalyst (3), the desired ROM/RCM product 325 was obtained, but in only 28% yield, due to the formation of the dimer 326 in 15% yield, together with recovered starting material in 47% yield. In an attempt
to improve upon this modest result, they found that exposure of dimer 326 to the Grubbs II catalyst (4) in ethylene atmosphere produced the desired product 325 in good yields. In practice, this tandem process was accomplished in one pot by sequential
treatment of 324 with the Grubbs I catalyst (3) in the presence of 2,6-dichloro-1,4-benzoquinone, followed by the treatment with
the Grubbs II catalyst (4) in atmospheric ethylene at 80 °C to provide the final product 325 in 81% overall yield. The formation of the oxirane ring was carried out in a highly
chemo- and stereoselective fashion by reaction of mCPBA with the cyclic olefinic bis-silyl ether derivative, obtained from 325 by reduction and bis-silylation of the resulting diol. The resulting epoxide 327 was then reacted with the Grubbs II catalyst (4) to obtain the macrocyclic derivative 328 in a notable 83% yield. The completion of the synthesis of clavilactone A (323) was efficiently accomplished in four additional steps, mainly functional groups
interconversions, through quinone 329 that corresponds to clavilactone B. In a similar synthetic sequence, they prepared
the originally proposed structure of clavilactone D (330) (Scheme [37]),[171] which led to its structural revision, identifying 331 as the correct structure.
Scheme 37 Total synthesis of clavilactone A, B, and D
A cascade process that involves an alkene metathesis with other types of reactions
has also been explored. As such, an application is the tandem CM/oxa-Michael cyclization
employed by Krische and Waldeck in the synthesis of the C2-symmetric natural product
cyanolide A (332) (Scheme [38]),[172] a potent molluscicidal agent against the water snail Biomphalaria glabrata, which actually is a vector of the human parasitic disease schistosomiasis. Thus,
treatment of diol 333 with H-G II catalyst (6) in the presence of pent-1-en-3-one gave compound 334 in 76% yield as a 10:1 mixture of diastereomers. This pyran derivative is the result
of an initial double CM reaction, followed by an oxa-Michael cyclization. In a second
CM round, compound 334 was treated with the Blechert catalyst 7 in an ethylene atmosphere to yield olefin 335 in 70% yield. A few more steps from 335 completed a concise total synthesis of cyanolide A (332) (Scheme [38]).
Scheme 38 Total synthesis of cyanolide A
Scheme 39 Total synthesis of decytospolide A and B
Another such application is the synthesis of the tetrahydropyran-containing natural
products decytospolide A (336) and B (337) by Kommu and co-workers (Scheme [39]),[173] which was rapidly achieved when the hydroxy olefin 338 was subjected to treatment by the H-G II catalyst (6) in the presence of pent-1-en-3-one to provide the pyrans 339 and 340 in 78% combined yield and as a 9:1 mixture of stereoisomers, in favor of the desired
2,6-cis-pyran 340. Removal of the benzyl group of 340 afforded decytospolide A (336), whose acetylation provided decytospolide B (337) (Scheme [39]).
An interesting approach towards the synthesis of isoquinoline-type alkaloids, such
as oxychelerythrine (341a), oxysanguinarine (341b), oxynitidine (341c), and oxyvicine (341d), has been described by Sutherland and co-workers[174] based on a tandem Overman rearrangement/RCM for rapid access to amino-substituted
1,4-dihydronaphthalene scaffolds, which represent key precursors for the synthesis
of the benzo[c]phenanthridine system found in these natural products (Scheme [40]). The one-pot Overman rearrangement/RCM was efficiently achieved after extensive
optimization, finding that when 342 was converted into the corresponding trichloroacetimidate, heated to 160 °C in the
presence of potassium carbonate and then submitted to the action of the Grubbs II
catalyst (4) at room temperature, the corresponding 1,4-dihydronaphthalene 343 was obtained in 81% yield. From this privileged compound, they completed the syntheses
of the alkaloids 341a–d, after aromatization of 343, followed by coupling with the corresponding 2-bromobenzoic acid derivatives, and
an intramolecular biaryl Heck coupling reaction, in a very highly efficient manner.
In the final Heck coupling (step h of the sequence), Sutherland and co-workers employed
the more stable Hermann–Beller palladacycle catalyst, given the very high temperature
(160 °C) required for the intramolecular coupling of their precursors 344a–d (Scheme [40]). A related tandem Overman rearrangement and ring-closing enyne metathesis, followed
by a Diels–Alder reaction was utilized by them in the synthesis of amino-substituted
indanes and tetralins.[175]
Scheme 40 Total synthesis of oxybenzo[c]phenanthridine alkaloids
3
Enyne Metathesis in Total Synthesis
As in previous sections, in this section we will focus on recent contributions recorded
in the last few years (2012 to early 2018), taking into account that the ring-closing
enyne-metathesis (RCEYM) reaction has been similarly covered in numerous reviews.[176]
3.1
Total Syntheses Based on a Ring-Closing Enyne-Metathesis Reaction
The synthetic value of an intramolecular metathesis reaction of an enyne (RCEYM) is
due to the versatility that is offered by the resulting cyclic dienes, which can be
utilized in subsequent transformations, such as cycloaddition reactions. On the other
hand, in contrast to other metathesis processes, the enyne metathesis can be promoted
by other metal catalysts, such as Pt2+, Pt4+, or Ir, or even Lewis acids, by a different mechanism, but with the same result.
Interestingly, in contrast to the alkene RCM, for the enyne RCM, the mode of the ring
closure can be different depending on the size of the cyclic system.[177] Thus, whereas in small- and medium-sized rings, the cyclization pathway goes through
an exo-mode that delivers the product type A, in a macrocyclization process, the ring closure can go through an endo-pathway, owing to the increased flexibility of the large ring system, that should
afford the product type B. A clear example of the formation of this class of type B compounds is the synthesis of 6-deoxyerythronolide B (345) by Krische and co-workers (Scheme [41]),[178] who prepared macrocyclic 347 in 89% yield when the enyne 346 was treated with H-G II catalyst (6) in an atmosphere of ethylene at 110 °C, with no detection of regioisomers. With
the formation of the 14-membered macrocycle, the completion of the synthesis of the
deoxyerythronolide B (345) was delineated through a synthetic sequence that included selective oxidation of
the exocyclic double bond, reduction of the resulting enone through a Ni-catalyzed
conjugate reduction, stereoselective methylation of the resulting ketone 348, and final removal of the protective groups via catalytic hydrogenation (Scheme [41]). Given the usual involvement of the enyne RCM products in subsequent transformations
of the resulting diene in a tandem process, the majority of cases found in the literature
correspond to this category and will be discussed later in Section 3.3.
Scheme 41 Different pathways of the ring-closing enyne metathesis and total synthesis of 6-deoxyerythronolide
B
3.2
Total Syntheses Based on an Enyne Cross-Metathesis Reaction
Several problems are associated with the selectivity of an intermolecular enyne metathesis,
for example alkene-alkene and alkyne-alkyne metathesis processes can compete with
the alkene-alkyne coupling, as well as the possible formation of geometric E/Z mixtures. Indeed, these problems are the likely reasons for the limited use of this
modality of metathesis versus the intramolecular variant. Due to the scarcity of examples
in which an enyne cross-metathesis is employed in the realm of total synthesis, it
is worth emphasizing those where the reaction has been successfully applied. Such
a case is the synthesis of amphidinolide P (349) by Diver and Jecs (Scheme [42]),[179] in which an enyne cross-metathesis allowed access to the main backbone of the natural
product. In practice, alkyne 350 and alkene 351 were coupled by treatment with the H-G II catalyst (6) at low temperature (–20 °C) to give diene 352 in 72% yield and as a 4.5:1 E/Z mixture of geometric isomers. The completion of the synthesis was achieved according
to the Williams protocol[180] to obtain pure amphidinolide P (349) in 38% overall yield. Surprisingly, when they attempted the synthesis of this natural
product via an enyne RCM, they failed to obtain any ring-closure product, either from
353 or from the more stable TMS derivative 354. However, when 353 was subjected to the Grubbs II catalyst (4) in an ethylene atmosphere, they could obtain, in almost quantitative yield, the
1,3-diene 355, which was envisioned as a suitable precursor for a final RCM step. In this case,
they observed that the RCM reaction of its corresponding silyl enol ether, compound
356, with the Grubbs II catalyst (4) proceeded in a better yield, compared with the direct RCM reaction of the diene
355, likely due to the introduction of additional rigidity in the system that would favor
the connection of the involved alkenes. The RCM product, obtained as a mixture of
the corresponding silyl enol ether and ketone, was transformed into amphidinolide
P (349) by reaction with HF·pyridine in 67% yield over two steps (Scheme [42]). This contribution joins the synthesis by Lee and co-workers based on an RCM approach.[181]
Scheme 42 Total synthesis of amphidinolide P
3.3
Enyne Metathesis in Tandem Reactions
Among the different variants of the metathesis reactions that employ enynes as starting
precursors, those that involve the enyne metathesis as part of a cascade process have
been the most exploited, due to their great potential for the generation of complex
polycyclic systems in a single step. Given the synthetic consequences of such processes,
enyne metathesis in cascade processes has occupied a strategic position in the synthesis
of natural products containing complex polycyclic systems and has been the topic of
several comprehensive reviews.[182] One finds two well-differentiated types of tandem processes involving enyne metathesis:
(a) Multiple intramolecular metathesis reactions in a programmed polyenyne system.
In this case, the polyenyne precursor is properly designed to trigger a highly orchestrated
cascade that ensures the correct regiochemical outcome, once the initial metal carbene
is formed in the correct position. (b) A tandem process consisting of an initial enyne
ring-closing metathesis, followed by a second reaction, usually a cycloaddition reaction,
of the resulting diene. In the first case (a), we find a relevant example in the formal
synthesis of englerin A (357) by Parker and Lee (Scheme [43]).[183] The intriguing structure of this natural product, in conjunction with its striking
pharmaceutical properties, has prompted intense research activity directed towards
its total synthesis, as well as the generation of analogues for medicinal chemistry
studies.
Scheme 43 Total synthesis of englerin A
Among the seminal and elegant contributions developed for the synthesis of this natural
product, the one by Parker and Lee was based on a ene-yne-ene metathesis cyclization
in a tandem process. In order to secure the initiation process in the correct direction,
they developed a relay metathesis cascade from precursor 358, in form of a diastereomeric mixture, which was readily available from geraniol in
seven steps. Thus, when compound 358 was treated with the Grubbs–Stewart catalyst (12), the expected tricyclic derivative 359 was obtained in 77% combined yield of both stereoisomers, which could be separated
by chromatographic methods. The power of this tandem metathesis reaction is reflected
by the fact that only the guaiadiene system was generated in a strict order of events
through ruthenium carbene species A and B due to the triggering effect exerted by the terminal allyl ether. The formation of
the oxygen bridge present in the natural product was then accomplished by means of
an oxymercuration and subsequent demercuration process that afforded the advanced
intermediate 360. The transformation of this compound into englerin A (357) has already been described by Echavarren in seven steps (Scheme [43]).[184]
Another representative and unique example is the synthesis of the complex tetracyclic
diterpenes belonging to the kempene family, such as the kempenone 361, which was isolated from the defense secretion of higher termites Nasutitermes octopolis. In contrast to the long multistep approaches reported by Paquette and Deslongchamps
for the construction of the tetracyclic core of these terpenes,[185] Metz and co-workers reported the preparation of this polycyclic system in one step
and in excellent yield from the dienyne 362 (Scheme [44]), which was prepared from the Wieland–Miescher ketone in 20 steps.[186] Thus, exposure of 362 to the Grubbs II catalyst (4) afforded the tetracyclic derivative 363 in an astonishing 97% yield. Whereas the reduction of the resulting tetracyclic ketone
363 with L-Selectride provided the alcohol 364 with the opposite stereochemistry to that of the natural product, reduction of the
bicyclic ketone precursor 362 provided the correct diastereomer 365 under the same reduction conditions. In a similar way as for 362, dienyne 365 was subjected to treatment by the Grubbs II catalyst (4) to obtain the tetracyclic alcohol 366 in almost quantitative yield (99%). From this compound, the synthesis of the kempenone
361 was achieved in nine additional steps (Scheme [44]).
Scheme 44 Total synthesis of kempenone
Yet another instructive example for this section is the elegant synthesis of (±)-morphine
(367) by Smith and co-workers (Scheme [45]),[187] in which the intricate pentacyclic system contained in this fascinating natural
product was generated utilizing a cascade ene-yne-ene ring-closing metathesis of the
highly functionalized benzofuran 368, prepared in just six steps from commercially available starting materials. In the
crucial metathesis event, the catalytic action of the H-G II catalyst (6) smoothly promoted a sequential cascade of events, initiated with an enyne-RCM reaction,
that generated the ruthenium alkylidene intermediate A, and continued with a final RCM reaction to produce tetracyclic compound 369. Despite this product being isolated in an impressive 94% yield, they decided to
continue with the synthetic sequence without isolation of the resulting tetracycle
369. The compound was then subjected to a 1,6-addition reaction by treatment with TFA
and then sodium carbonate. As a result, a 10:1 mixture of compounds corresponding
to neopinone (370) and codeinone (371) was obtained. The treatment of this mixture with HCl, followed by NaOH workup, allowed
for the isomerization of neopinone (370) to codeinone (371). The reduction of 371 with sodium borohydride afforded codeine (372) as a single diastereomer in 65% overall yield from 368. Finally, demethylation of codeine with boron tribromide yielded racemic morphine
(367) in 86% yield (Scheme [45]).
Scheme 45 Total synthesis of (±)-morphine
Scheme 46 Total synthesis of virgidivarine and virgiboidine
An interesting example of the synthetic application of domino metathesis in the field
of the alkaloid-type natural products is the recent syntheses of virgidivarine (373) and virgiboidine (374), which are piperidine and piperidino-quinolizidine alkaloids isolated from the leaves
of African leguminosae Virgilia divaricata and Virgilia oroboides and whose biological activities have not been studied. Blechert and co-workers (Scheme
[46])[188] planned the construction of the common dipiperidine core of these natural products
via a ene-ene-yne ring-rearrangement metathesis (RRM) from the dienic system represented
by compound 375, foreseeing an initial ROM process of the cyclopentene unit, followed by a double
alkene RCM-enyne RCM process of the resulting ROM intermediate. Thus, when precursor
375, prepared from cis-cyclopent-2-ene-1,4-diol via enzymatic desymmetrization, was exposed to the H-G II
catalyst (6) under ethylene atmosphere, compound 376 was obtained in a remarkable 83% yield, according to the expected mechanistic course
delineated by them. Combined with the tandem metathesis reactions, the resulting metathesis
product was transformed into the aldehyde 377 by oxidative treatment in 33% overall yield in one pot. From this aldehyde, both
alkaloids were prepared, requiring ten steps to synthesize virgidivarine (373), followed by the synthesis of virgiboidine (374), after a final lactamization reaction of 373 with EDC (Scheme [46]).
With regards to recent examples, special attention is deserved for the synthesis of
the anticancer alkaloids flueggine A (378) and virosine B (379) by Li, Yang, and co-workers (Scheme [47]).[189] These compounds, belonging to the securinega family, possess unique structural features
and interesting biological properties that cover a wide range of activities. The synthesis
of these compounds was based on an elegant RRCM of a dienyne system for the controlled
construction of the core structure represented by norsecurinine (385). Thus, their syntheses were envisioned to proceed via a 1,3-dipolar cycloaddition
between norsecurinine (385) and nitrone 386, which would be prepared from norsecurinine (385) by oxidative treatment with mCPBA. In a similar strategy towards the related alkaloid securinine, Honda and co-workers
(Scheme [47])[190] attempted its synthesis through a tandem RCM of the dienyne 380. However, when this precursor was reacted with different ruthenium catalysts, such
as the Grela–Grubbs–Hoveyda catalyst (10), the desired γ-lactone 381 was not detected, obtaining instead the δ-lactone 382 in 69% yield. The preference of the catalyst to react with the butenyl group rather
than the acrylate moiety, through intermediate A, could explain the outcome of this reaction. In an attempt to favor the formation
of the required ruthenium complex at the less reactive alkene, Li, Yang, and co-workers[189] proposed the trienyne 383 (Scheme [47]) as a suitable precursor to promote the preferential formation of intermediate C through the initiation of the reaction at the less sterically encumbered terminal
olefin. In fact, when compound 383 was subjected to the action of the Zhan-1 b catalyst (11), among others, compound 384 was obtained in a reproducible and acceptable 64% yield. The completion of the synthesis
of norsecurinine was undertaken in three additional steps, allylic bromination, Boc
deprotection, and a nucleophilic cyclization under basic conditions. After some additional
experimentation, they found that this transformation could be achieved in only one
step by exposure of the resulting bromide to the acid AgSbF6 in an improved 85% yield. Finally, the preparation of the nitrone 386 from norsecurinine (385), according to the Magnus protocol,[191] was followed by a cycloaddition reaction with 385 to obtain natural product flueggine A (378) in 77% yield. In a related approach, they also prepared virosaine B (379) by an intramolecular cycloaddition of nitrone 388, prepared from the isomer allonorsecurinine (387) in three steps (Scheme [47]).
Scheme 47 Total synthesis of flueggine A and virosaine B
In the second group of the cascade reactions involving enyne-RCM processes, we wish
to highlight those that combine the metathesis reaction with a Diels–Alder cycloaddition
of the resulting diene as a way for the construction of a bicyclic system in a rapid
and efficient fashion. Within this category of tandem reactions, the synthesis of
the norsesquiterpene-type alkaloids, such as the anticancer alkaloid 389, which were isolated from the culture of mushroom-forming fungus Flammulina velutipes, was accomplished using a tandem enyne-RCM/Diels–Alder/aromatization sequence in
one-pot by Reddy and co-workers (Scheme [48]).[192] Thus, the treatment of the acyclic precursor 390 with the Grubbs I catalyst (3) in toluene at 50 °C was followed by a Diels–Alder reaction with dimethyl acetylenedicarboxylate
and then, treatment with DDQ of the resulting Diels–Alder adduct 391, which was not isolated, to provide the final compound 392 in a moderate 42% overall yield. The elaboration of this product towards the final
alkaloid 389 was accomplished without issues in three additional steps in 49% overall yield (Scheme
[48]).
Scheme 48 Total synthesis of the norsesquiterpene alkaloid
Figure 7 Common catalysts used for alkyne-metathesis reactions
4
Alkyne Metathesis in Total Synthesis
4.1
Total Syntheses Based on a Ring-Closing Alkyne-Metathesis Reaction
The wide range of reactivities that alkynes possess, compared with alkenes, imparts
alkyne metathesis as a very powerful tool with a broader field of synthetic opportunities.
Thus, some advantages of alkynes over alkenes include the stereocontrolled partial
reduction of the triple bond, which allows stereoselective access to the E- or Z-alkene, and its excellent reactivity against electrophilic metals such as Ag, Pt,
or Hg, which allows for further functionalization. On the other hand, the catalysts
required for the alkyne-metathesis reactions are completely different. From the first
Mortreux system developed in 1974,[193] the evolution towards more efficient, stable and functional group tolerant catalysts,
developed first in 1982 by Schrock (393)[194] and later in 1999 by Fürstner (394)[195] (Figure [7]), has propelled the alkyne-metathesis reaction to become an excellent synthetic
tool for the generation of macrocyclic structures as applied to the synthesis of natural
products and bioactive compounds. Since the first example of an alkyne ring-closing
metathesis in 1999 by Fürstner,[196] we can find in the literature many applications of this type of reaction in the
preparation of complex macrocyclic natural products, some of the which have been linked
to the development of even more efficient and stable catalysts (395–399) since 2011.[197] As for the catalysts used in alkene metathesis, the catalysts for alkyne metathesis
are precatalysts, which are activated to form the active species responsible for catalytic
action. In fact, precatalysts 394 or 399 are activated by reaction with CH2Cl2 to form in situ the corresponding monochloro derivatives, which are the catalytically
active species.[198] One of the main architects of the responsible for the spectacular irruption of this
reaction in modern organic synthesis has been Fürstner, who by both his seminal contributions
in total synthesis and his work on the development of new catalysts has reached spectacular
progress and promoted the field. During his prolific and impressive research career,
many reviews have been reported about this reaction covering the main contributions
of the Fürstner group as well as of other groups.[199] The contributions reported in the period 2015–2017, however, have not been covered
and thus will be reviewed in this section.
Thus, selected examples are illustrated in Scheme [49], in which the natural products lactimidomycin (400),[200] mandelalide A (401),[201] amphidinolide F (402),[202] brefeldin A (403),[203] leidodermatolide (404),[204] tulearin A (405),[205] or brominated 4-pyrone-type natural products[206] (e.g., 406) were efficiently synthesized from their corresponding acyclic dialkynes which gave
the corresponding alkyne macrocycles 407–412 in good to excellent yields. In the depicted representative cases, the alkynes were
reduced to the corresponding E-alkenes (cases of natural products 400, 403, and 405), Z-alkenes (cases of 401 and 404), or even, not transformed, such as the cases of the 4-pyrone 406. For the particular case of amphidinolide F (402), the alkyne was subjected to a hydration reaction to obtain a ketone under platinum
catalysis. Similarly relevant are the syntheses of WF-1360F (413),[207] cruentaren A (31),[208] and haliclonin A (414),[209] reported by the Altmann, Barrett, and Huang groups, respectively, in which after
the alkyne macrocycle formation via a RCAM reaction, the resulting macrocyclic alkynes
415–417 were reduced to the E-olefin in the case of 415 or to the Z-olefins by means of a Lindlar catalyst mediated hydrogenation for 416 and 417 (Scheme [50]).
Scheme 49 Selected natural products synthesized via a RCAM reaction by the Fürstner Group (2013–early
2018)
Scheme 50 Selected natural products synthesized via a RCAM reaction by other research groups
(2013–early 2018)
Scheme 51 Selected natural products synthesized via RCAM/hydrostannation reactions by the Fürstner
group
More intriguing and challenging are the cases in which the introduction of a methyl
group into the alkyne is required to be transformed into a methyl-branched alkene
in a regio- and stereocontrolled manner. Such cases can be found in the synthesis
of the natural products 5,6-dihydrocineromycin B (418), disciformycin B (419), and nannocystin Ax (420) by the Fürstner group. To this aim, the recent methodology developed by this group,
based on the ruthenium-catalyzed trans-hydrostannation of alkynes, was envisioned as a solution to this synthetic problem
given the excellent levels of regioselectivity exhibited by this reaction when applied
to unprotected propargyl alcohol derivatives. Thus, in the case of 5,6-dihydrocineromycin
B (418) (Scheme [51]),[210] when alkyne 421, smoothly obtained from the corresponding acyclic dialkyne by reaction with the molybdenum
alkylidyne complex 397, was desilylated by reaction with HF, the resulting propargyl alcohol was treated
with Bu3SnH in the presence of the ruthenium catalyst [Cp*RuCl2]n. As a result, the corresponding α-alkenylstannane was obtained as a single isomer.
Having introduced the stannyl group, a Stille reaction was carried out by exposure
to copper thiophene-2-carboxylate (CuTC), [Ph2PO2][NBu4], and Pd(PPh3)4 in the presence of methyl iodide to afford the final natural product 418 in excellent yield. It is important to highlight the importance of the order of addition
and stoichiometry of the reactants in this reaction to avoid undesired protodestannation
processes. In a similar situation, the macrocyclic alkyne 422 (Scheme [51]),[211] obtained in excellent yield from its corresponding dialkyne precursor by reaction
with precatalyst 399, required the installation of the methyl group for the synthesis of disciformycin
A and B, interesting macrolides-type antibiotics with considerable activity against
resistant Gram-positive bacteria. As the previous case, the transformation of compound
422 into its hydroxy derivative was followed by a rapid trans-hydrostannation reaction. In contrast to the previous case, as to other numerous
examples carried out by the Fürstner group, on this occasion a 3:1 mixture of stannyl
regioisomers, in favor of 423, was obtained, likely due to the competing interactions of the hydroxy group, via
hydrogen bonding, with the ruthenium catalyst and with the neighboring ester carbonyl
group. In addition to this lack of regioselectivity, the observation that the stannyl
derivative was quite sensitive forced them to continue with the crude product mixture
to the following step, which consisted of a methylation reaction. Once again, this
reaction proved to be sluggish, as a mixture of products was obtained under conventional
conditions.
In light of these unsatisfactory results, they decided to use stoichiometric amounts
of the methyl donor [(cod)Pd(Me)Cl] to obtain the desired product 424 in 20% overall yield after four steps from 422. Final glycosidation and global deprotection provided the elusive disciformycin B
(420) in the first instance, and then disciformycin A after an isomerization reaction.
Similarly challenging proved to be the synthesis of nannocystin Ax (420) (Scheme [51]),[212] a cytotoxic cyclodepsipeptide isolated from myxobacteria. After a successful alkyne
RCM reaction that provided macrocyclic alkyne 425 in 66% yield by the action of catalyst 395, the hydroxy-directed hydrostannation was achieved by sequential desilylation and
treatment with Bu3SnH, in the presence of catalytic amounts of [Cp*RuCl2]n, to obtain the corresponding stannane 426 as a single regio- and stereoisomer. The methylation was then achieved according
to previous conditions to provide the desired methyl derivative in a reasonably good
yield (70%). The methylation of the hydroxy group proved to be similarly challenging,
finding, after extensive experimentation, that the methyl ester 427 provided a proper source of electrophilic methyl by its in situ generation in a gold-catalyzed
cyclization. Final reductive cleavage of the phenacyl group furnished the targeted
natural product 420 (Scheme [51]).
Beyond the transformation of the alkyne into an alkene or even a methyl-substituted
alkene, as discussed vide supra, the reactivity of the alkyne functional group allows the construction of a heterocyclic
ring, providing further extraordinary opportunities to access more complex structural
systems compared to the alkene chemistry.
These post-metathetic transformations have been similarly explored by the Fürstner
group in the synthesis of a plethora of natural products. In this set of new synthetic
opportunities, the syntheses of enigmazole A (428) (Scheme [52]),[213] polycavernoside A (429) (Scheme [52]),[214] kendomycin (430) (Scheme [53]),[215] and lytharanidine (431) (Scheme [53])[216] represent excellent examples of this synthetic potential. In all these cases, the
macrocyclic alkynes 432–435 were efficiently synthesized from their corresponding acyclic dialkynes. From these
alkynes, in the cases of enigmazole A, polycavernoside A, and kendomycin, a transannular
pyran (for enigmazole A) and a furan ring (for polycavernoside A and kendomycin) were
efficiently constructed by means of activation of the alkyne with gold-based catalysts
436–438
[217] to provide compounds 439–441. In the cases of polycavernoside A and kendomycin, the syntheses were formal, having
been described their total synthesis by Lee and Woo[218] and Mulzer and co-workers[219] from advanced precursors 440 and 442, respectively. On the other hand, for the synthesis of lytharanidine (431), formation of a piperidine ring was required, which was achieved by a sequence of
redox isomerization, followed by transannular aza-Michael addition from macrocyclic
alkyne 435 (Schemes 52 and 53).
Scheme 52 Selected natural products synthesized via a RCAM/gold catalysis by the Fürstner group
Scheme 53 Selected natural products synthesized via a RCAM/gold catalysis by the Fürstner group
Scheme 54 Total synthesis of ivorenolide B
An intriguing last case is the synthesis of ivorenolide B (443) by the Fürstner group based upon a macrocyclization of a diterminal diyne to prepare
the 1,3-diyne system found in this unprecedented natural product (Scheme [54]).[220] Ivorenolide B belongs to a family of macrolides isolated from the stem bark of Khaya ivorensis, which has been reported to inhibit concanavalin A induced T-cell and lipopolysaccharide-induced
B cell proliferations. This noteworthy biological activity confers it a clinical application
as an immunosuppressive agent comparable to cyclosporine A. The synthesis by the Fürstner
group was delineated according to a RCAM reaction of the diterminal acyclic diyne
444, which by treatment with catalyst 395 furnished in a remarkable 82% yield the desired [17]-membered ring 445 with the embedded 1,3-diyne unit and no detection of side products derived from ring
contraction or oligomerization processes. The subsequent epoxidation, performed in
a regio- and stereoselective manner with mCPBA, followed by silyl ether deprotection provided the targeted ivorenolide B (443) (Scheme [54]). This synthesis improved notably by Yue and co-workers,[221] who, based on a RCM reaction, provided the macrocyclic lactone 446 as a 1.5:1 mixture with the E-olefin as the major isomer. Despite this stereochemical problem, this synthesis allowed
the absolute configuration of the natural product to be confirmed (Scheme [54]).
4.2
Other Types of Alkyne-Metathesis Reactions
To conclude this review, it is necessary to mention that, as in alkene metathesis
reactions, we can find also examples of the use of the alkyne cross-metathesis[222] or tethered RCAM reactions.[223] Despite the power and scope that the alkyne metathesis demonstrates, utilization
of these alkyne metathesis variants in the context of total synthesis is not well
represented, with most examples being found mainly in polymerization reactions.[224] Nevertheless, given the potential and efficiency of the methodology, there is no
doubt that the trend will be an increase in the use of this chemistry in the construction
of natural products.
