Key words cross-metathesis - cross-coupling - stereocontrolled synthesis - conjugated diene
- rhizopodin
Rhizopodin was isolated by Höfle and Reichenbach from the myxobacterium Myxococcus stipitatus and was assigned as a monomeric lactone in 1993.[1 ] Its structure and absolute stereochemistry were recently revised as shown in Scheme
[1 ].[2 ]
[3 ] Rhizopodin exhibits significant biological properties including potent cytostatic
activity in the low nanomolar range against a range of tumor cell lines.[1–3 ] The distinctive structural features and biological activities, together with our
interest in macrocyclic marine natural products[4 ] prompted us to undertake studies on the synthesis of rhizopodin. Recently, various
synthetic approaches toward the synthesis of rhizopodin have been reported.[5 ] The syntheses of monorhizopodin and 16-epi -monorhizopodin were achieved by Nicolaou and co-workers in 2011[5e ] and, since then, two total syntheses of rhizopodin have been reported.[5g ]
[j ]
So far, total syntheses of the macrocycle of rhizopodin have employed either intramolecular
Suzuki coupling reaction or macrolactonization.[5g ]
[h ]
[i ]
[j ] An alternative approach to the macrocyclizations was sought and we opted to close
the macrocyclic core by ene-diene cross-metathesis[6 ] as shown in our retrosynthetic plan (Scheme [1 ]). To test the feasibility of the key ene-diene cross-metathesis step of our designed
strategy toward rhizopodin, a model study based on the construction of the C1–C15
fragment (1 ) of rhizopodin was undertaken. Herein we detail two synthetic approaches to fragment
1 .
Scheme 1 Retrosynthetic analysis
Retrosynthetic analysis of 1 led us to disconnect between positions C8 and C9, which imposed the construction
of the conjugated diene through cross-metathesis of fragments 2 and 3 . Alternatively, a Suzuki cross-coupling of vinyl boronate 4 with vinyl iodide 5 was envisioned to deliver diene 1 (Scheme [1 ]). Oxazole-containing fragments 2 and 4 were planned to originate from the common precursor 10 , which, in turn, would be prepared from the known methyl-2-(chloromethyl)oxazole-4-carboxylate
(7 ; Scheme [2 ]).
The synthesis of fragments 2 and 4 is outlined in Scheme [2 ]. Oxazole 7 was obtained from commercially available 2,2-dichloronitrile (6 ) by using a known sequence.[7 ] Reaction of 7 with sodium acetate in the presence of acetic acid and acetic anhydride and treatment
of the resultant acetate derivative with potassium carbonate and methanol afforded
the corresponding alcohol 8 in 61% yield over two steps. After protection of the primary alcohol as its TBS ether,
the methyl ester was reduced with DIBAL-H in THF to give alcohol 9 in 84% yield. This route is operationally convenient and proceeds well on large scale
(>35 g of 9 was obtained). It should be mentioned that alcohol 9 could be obtained by a reported procedure;[8 ] however, in our hands, we were unable to reproduce the reaction on large scales.
Swern oxidation[9 ] of the primary alcohol of 9 , followed by Keck allylation[10 ] of the resulting aldehyde, provided homoallylic alcohol 10 in 72% yield with >97% enantiomeric excess, as measured on its Mosher ester.[11 ] The absolute stereochemistry at the newly created stereogenic center was also assigned
at this point by synthesis and comparison of the 1 H NMR spectra of its Mosher derivatives.[12 ] Thus, homoallylic alcohol 10 was reacted with both (S )- and (R )-α-methoxy-α-(trifluoromethyl)phenylacetic acid to generate diastereomeric (S )- or (R )-Mosher esters 12 and 13 , respectively (Table [1 ]). Subtraction of the chemical shifts of the protons of (R )-Mosher ester 13 from those of (S )-Mosher ester 12 in the vicinity of the ester-bearing stereocenter then provides differences (Δδ ), the signs of which are used to assign the configuration of the stereocenter. The
signs of the Δδ are shown in Table [1 ], and the absolute stereochemistry at C11 was elucidated to be the (S )-configuration.
Scheme 2 Preparation of intermediates 2 and 4 . Reagents and conditions : (i) NaOAc, HOAc-Ac2 O; (ii) K2 CO3 , MeOH; (iii) TBSCl, imidazole, CH2 Cl2 ; (iv) DIBAL-H, THF; (v) (COCl)2 , DMSO, Et3 N, CH2 Cl2 , –78 °C; (vi) (S )-BINOL, Ti(Oi -Pr)4 , allyltributylstannane, CH2 Cl2 , –20 °C, 72 h; (vii) NaH, MeI, THF; (viii) 11 , Hoveyda–Grubbs II catalyst, toluene, 80 °C.
Table 1 Stereochemical Assignment of 10
a
Hydrogen
δ
S (S -MTPA ester)
δ
R
(R -MTPA ester)
δ
S
–δ
R
TBS(t -Bu)
0.910
0.901
+0.009
TBS(Me)
0.104
0.088
+0.016
15
4.742
4.701
+0.042
13
7.574
7.409
+0.165
11
6.050
6.077
–
10
2.745
2.790
–0.045
9
5.634
5.759
–0.125
8
5.082
5.171
–0.089
a Reaction conditions: (i) oxallyl chloride, DMF, CH2 Cl2 , (S )-MTPA; then Et3 N, DMAP, CH2 Cl2 , 10 , 85%; (ii) oxallyl chloride, DMF, CH2 Cl2 , (R )-MTPA; then Et3 N, DMAP, CH2 Cl2 , 10 , 91%.
O -Methylation of the homoallylic alcohol 10 by using iodomethane and sodium hydride in THF provided 2 in 84% yield. Olefin cross-metathesis between 2 and vinyl pinacol boronate 11 under the influence of the Hoveyda–Grubbs II catalyst furnished vinyl boronate 4 in 65 % yield (E /Z = 10:1; Scheme [2 ]).[13 ]
The preparation of coupling partners 3 and 5 commenced from methyl ester 14 , which was prepared according to conditions described by Paterson[14 ] (Scheme [3 ]). Methyl ester 14 was reduced to aldehyde 15 in 94% yield by using DIBAL-H in dichloromethane. Subsequent treatment of 15 with β -(+)-allyldiisopinocampheylborane according to Brown et al.[15 ] provided a 10:1 mixture of syn - and anti -monomethylated diols, favoring the desired isomer, which was protected as its TBS
ether 16 (55% yield over two steps). Oxidative cleavage of the terminal olefin by using the
Sharpless protocol,[16 ] and esterification of the resultant acid with trimethylsilyldiazomethane[17 ] provided methyl ester 17 in 63% yield over two steps. Selective removal of the primary TBS group in the presence
of the secondary TBS group under mild acidic conditions provided alcohol 18 in 88% yield. Oxidation of the primary alcohol in 18 by using the Dess−Martin periodinane reagent[18 ] buffered with sodium bicarbonate afforded the corresponding aldehyde, which served
as the precursor leading to both 3 and 5 . Thus, a Horner–Wadsworth–Emmons olefination between the diethyl allylphosphonate
and the aldehyde derived from 18 successfully led to the required diene 3 in 53% yield and good E /Z selectivity (15:1). In parallel, treatment of the aldehyde with iodoform and chromous
chloride in THF[19 ] gave vinyl iodide 5 in 63% overall yield with greater than 10:1 E /Z selectivity.
Scheme 3 Preparation of intermediates 3 and 5 . Reagents and conditions : (i) DIBAL-H, CH2 Cl2 , –78 °C; (ii) (+)-Ipc2 BOMe, allylmagnesium bromide, Et2 O, 0 to –78 °C; then 15 ; (iii) TBSCl, Et3 N, DMAP, CH2 Cl2 ; (iv) RuCl3 , NaIO4 , CCl4 –MeCN–H2 O; (v) TMSCHN2 , MeOH; (vi) (–)-CSA, CH2 Cl2 –MeOH; (vii) Dess–Martin periodinane, NaHCO3 , CH2 Cl2 ; (viii) diethyl allylphosphonate, n BuLi, HMPA, THF, –78 °C, then aldehyde; (ix) CrCl2 , CHI3 , THF.
With the four requisite fragments 2 –5 in hand, the stage was set to test the palladium-catalyzed cross-coupling reaction
(Table [2 ]) and ene-diene cross-metathesis (Table [3 ]). First, we explored the Suzuki coupling reaction[20 ] of vinyl boronate 4 and vinyl iodide 5 under various conditions. Initially the widely used protocol was employed using [Pd(Ph3 P)4 ]-Ph3 As and Cs2 CO3 in THF. Unfortunately, the desired product 1 was obtained in only 35% isolated yield (Table [2 ], entry 1). When the catalyst was switched to [Pd(dppf)2 Cl2 ], diene 1 was isolated in a much improved yield (52%; entry 2). Further optimization of the
catalytic systems led to the identification of a remarkably simple protocol, in which
vinyl boronate 4 and vinyl iodide 5 were exposed to [Pd(Ph3 P)4 ] and TlOEt in THF–H2 O (v/v 4:1) to furnish 1 in 73% isolated yield (entry 3).[21 ]
Table 2 Suzuki Coupling for the Synthesis of Diene 1
a
Entry
Ratio 4 /5
Catalyst/ligand
Solvent
Base
Yield (%)
1
1.3:1.0
[Pd(Ph3 P)4 ]/Ph3 As
THF
Cs2 CO3
35
2
1.3:1.0
[Pd(dppf)2 Cl2 ]/Ph3 As
THF
Cs2 CO3
52
3
1.3:1.0
[Pd(Ph3 P)4 ]
THF–H2 O
TlOEt
73
a Reaction conditions: See table and text for details.
We then examined the key ene-diene cross-metathesis reaction under several conditions
by varying the solvent, catalyst, and temperature. As shown in Table [3 ], initial cross-metathesis of alkene 2 and diene 3 at 80 °C in toluene with Grubbs I catalyst provided no conversion (entry 1). Furthermore,
attempts to mediate the cross-metathesis with 10 mol% of either Grubbs II catalyst
or Hoveyda–Grubbs II catalyst in dichloromethane at reflux also failed to provide
any detectable quantities of diene 1 . Fortunately, performing this reaction at 60 °C in toluene with Grubbs II catalyst
or Hoveyda–Grubbs II catalyst, afforded 1 in 24% and 40% yield, respectively (entries 4 and 5). Furthermore, by increasing
the reaction temperature to 80 °C in toluene in the presence of Hoveyda–Grubbs II
catalyst, the yield improved considerably, and diene 1 could be isolated in 51% yield[21 ] as the (E ,E )-alkene, the conformation of which was determined by 1 H NMR spectroscopic analysis. Considering the thermal stability of both starting material
and catalyst, attempts to further improve the yield by the use of elevated temperatures
were not conducted.
Table 3 Cross-Metathesis of Alkene 2 and Diene 3
a
Entry
Catalyst (10 mol%)
Solvent
Temp (°C)
Yield (%)b
1
Grubbs I
toluene
80
–
2
Grubbs II
CH2 Cl2
40
–
3
Hoveyda–Grubbs II
CH2 Cl2
40
–
4
Grubbs II
toluene
60
24
5
Hoveyda–Grubbs II
toluene
60
40
6
Hoveyda–Grubbs II
toluene
80
51
a Reaction conditions: See table and text for details.
b Yield based on recovered 3 .
In summary, the C1–C15 fragment of rhizopodin was synthesized by either Suzuki coupling
reaction of vinyl iodide and vinyl boronate or by cross-metathesis of a terminal olefin
and a diene adduct in the presence of Hoveyda–Grubbs II catalyst. This study demonstrates
the effectiveness of an ene-diene cross-metathesis approach to diene 1 and served as a model study for the total synthesis of rhizopodin based on ene-diene
cross-metathesis strategy. Further efforts directed toward the asymmetric total synthesis
of rhizopodin and its analogues are underway and will be reported in due course.
