methylcarbapenem 8 in low yield (10% yield after a quick
nitromethane HOMO corresponded well with the LUMO of
the phosphorus atom in 13 but did not match the LUMO at
C2. This is consistent with the principal attack being on the
phosphorus of the phosphate. In contrast, the LUMO at C2
of triflate 12 is better matched with the HOMO of aci-
nitromethane than the sulfur atom’s, and therefore the
preferred attack is on the C-2 carbon rather than the sulfur
of the triflate.
purification through a short silica column). The structure of
1
13
8
was confirmed by H and C NMR (CDCl
3
) [AB quartet
at δ 5.97 and 5.10 for the nitromethylene protons] and by
CI mass spectra (NH
3
) with observed molecular ions at 520
+
+ 10
(MH ) and 537 (M + NH
4
).
In an effort to improve the yield, we studied two other
substrates in the nitromethane addition reaction. Starting from
enol phosphate 5, the desired product 7 was again formed
in low yield with keto-ester 9 as the major product as judged
by LC and NMR of the reaction mixture after workup
(Scheme 2). Further attempts were carried out in the presence
Scheme 2
Figure 1. Frontier orbitals.
With 8 in hand, we next examined the Pd-catalyzed allylic
14
coupling with hydroxyethylnaphthosultam 14. Initial studies
with isolated nitromethyl compound 8 under biphasic condi-
3 2 3
tions (toluene/THF, aqueous NaHCO , Pd(OAc) , (EtO) P)
of cosolvents, DMPU or HMPA.11 In all cases, keto-ester
, presumably arising from attack on the phosphorus, is the
major reaction product and the desired product 7 was the
produced low yields of desired product 15, along with
substantial quantities of p-nitrobenzyl alcohol. Given the
instability of the isolated nitromethyl intermediate, our
attention was then focused on developing a through process
from the enol triflate. Thus, treatment of enol triflate 6 in
9
1
2
minor product, formed in respective ratios of 3-5:1.
Significant improvement was obtained by starting with
triflate 6 and DMPU as the cosolvent with nitromethane.
After workup, the unstable oily product was partially purified
by being filtered through a silica pad. The estimated yield
of 8 was 60-70% based on LC and low-temperature NMR
analysis. The ratio of nitromethyl product 8 to keto-ester
1
3 2
:1 DMPU/CH NO at -20 °C with 6 equiv of TMG
followed by aging for 1-2 h afforded complete consumption
of starting material (Scheme 3). Subsequent partial neutral-
1
0 ranged from 5 to 10:112 depending on the amounts of
Scheme 3
base and nitromethane used.
To rationalize the observed reversal of chemoselectivity
between the two substrates and gain insight into the electronic
nature of these reacting species, we have performed molec-
ular orbital calculations on model compounds 11, 12, and
1
3. Frontier orbitals for aci-nitromethane and two carbap-
1
3
enems were generated using PC SPARTAN program at
the 3-21G* level (Figure 1). We found that the aci-
ization with MsOH (4 equiv), addition of naphthosultam 14
(
2 3
1 equiv), Pd(OAc) (5%), and (EtO) P (15%) and aging at
(
10) 8: 1H NMR (CDCl3) δ 8.19 (d, J ) 8.8 Hz, 2H), 7.63 (d, J ) 8.8
Hz, 2H), 5.97 (AB d, J ) 14.9 Hz, 1H), 5.43 (AB d, J ) 13.9 Hz, 1H),
.26 (AB d, J ) 13.9 Hz, 1H), 5.10 (AB d, J ) 14.9 Hz, 1H), 4.37 (dd, J
10.7, 3.4 Hz, 1H), 4.26 (dd, J ) 6.0, 5.2 Hz, 1H), 3.35 (dd, J ) 4.8, 3.4
rt for 1-2 h led to complete consumption of 8 to afford the
desired product 15. The overall assay yield of 15 is 34% for
the two steps. Workup with pH 7 buffer provided a 97%
recovery of product with good rejection of unreacted 14.
Thus, a sequential nitromethane conjugate addition reaction
to triflate 6 and the Pd(0)-catalyzed coupling of the resulting
nitromethyl 8 with hydroxyethylnaphthosultam 14 was
demonstrated.
5
)
Hz, 1H), 3.29 (m, 1H), 1.21 (d, J ) 6.0 Hz, 3H), 1.18 (d, J ) 8.0 Hz, 3 H),
13
0
.84 (s, 9H), 0.07 (s, 3H), 0.06, (s, 3H); C NMR (CDCl3) δ 174.6 (CON),
1
(
60.2 (CO2), 147.7, 142.1, 136.7, 132.3, 128.1 (PNB), 123.9 (PNB), 70.7
CH2NO2), 65.9 (CH2(PNB)), 65.3 (C8), 61.2 (C6), 55.5 (C5), 40.7 (C1),
2
5
5.6 (tBu), 22.1 (C9), 15.1 (C1a), -4.3 (SiMe), -5.1 (SiMe); MS (CI),
+
+
20 (MH ) and 537 (M + NH4 ).
11) Previous studies indicated that HMPA or DMPU can stabilize the
carbapenem: see ref 5.
(
(
12) Based on LC area % at 260 nm.
(13) Available from Wavefunction, Inc., 18401 Von Karman Ave., Suite
(14) Humphrey, G. R.; Miller, R. A.; Lieberman, D. R. W.O. Patent
9851677, 1998.
3
70, Irvine, CA 92612.
1784
Org. Lett., Vol. 1, No. 11, 1999