72 J . Org. Chem., Vol. 64, No. 1, 1999
Boger et al.
Sch em e 2
cine with 12, were accumulated. Without optimization,
subjection of 15 to macrocyclization with K2CO3/CaCO3
(5.0 equiv/7.5 equiv, 2 wt equiv of 4 Å MS, 0.008 M DMF,
48 °C, 24 h) provided a 4:1 mixture of the two separable
atropisomers of 16 (43%) in which the unnatural M-
isomer was favored (Scheme 2), and the closure proceeded
at a rate analogous to that observed with 4. The com-
parison of 4 and its minor C8 diastereomer19 with the
two atropisomers of 16 confirmed that they are distin-
guishable and that the preparations of 4 proceeded
without detectable C14 epimerization.20 Attempts to
cyclize 15 with CsF (5 equiv, 0.005 M DMF or DMSO,
25 °C) provided a much more polar set of atropisomers
in yields as high as 75% containing isomers which were
easily distinguishable from 4 or 16. These possessed a
higher molecular weight, indicative of further substrate
reactions, and although their identification was not
pursued, more detailed studies with 4 implicated C8
oxidation.21
enlisting the corresponding nitrile. A (4,4’-dimethoxy-
diphenyl)methyl (Ddm)22 protected L-Asn was also ex-
amined in light of its use in related efforts8,14 to assess
whether it might favorably impact the macrocyclization
reaction. In particular, we were interested in establishing
whether the liberated fluoride might prove sufficiently
basic to promote â-elimination through R-deprotonation
of the nitrile, thus lowering the apparent effectiveness
of our original approach. Thus, coupling of N-CBZ-N’-
Ddm-(S)-Asn22 (17) with (R)-8 cleanly provided the
dipeptide 18 (92%) as a single detectable diastereomer
(Scheme 3). N-CBZ deprotection (H2, Pd/C, 100%) fol-
lowed by coupling of the free amine 19 with 12 (2.2 equiv
of EDCI, 1.2 equiv of HOBt, DMF, 0-4 °C, 15 h, 73%)
provided the tripeptide 20 as a separable 1.8:1 mixture
of diastereomers without optimization.17 Under compa-
rable conditions, the L-(â-cyano)alanine-containing dipep-
tide provided a more favorable 6:1 ratio of C14 diaster-
eomers.11 Macrocyclization upon treatment with K2CO3/
CaCO3 (5.0 equiv/7.7 equiv, 2 wt equiv of 4 Å MS, 0.008
M DMF, 45 °C, 47 h) provided a mixture of the two
atropisomers of 2120 in a yield (47% versus 59%) and
atropisomer diastereoselectivity (1:1.3 versus 1:1.3 M:P)
comparable, but not superior, to the substrate bearing a
nitrile. Analogous to observations made with 16 and in
our prior efforts,11 attempted closure of 20, enlisting CsF
Alter n a tive L-Asn Resid u es. In our prior studies, the
L-Asn residue in the DE ring system was protected
(19) Trace amounts (<7-8%) of a C8 diastereomer were detected
and isolated when the macrocyclization leading to 4 was prematurely
quenched. The diastereomer, not detected in the closure of 23, rapidly
and completely epimerized to 24 upon treatment with base. Thus,
treatment of (P)-30 or (P)-24 with 2 equiv of DBU (THF-d8, 25 °C) or
3 equiv of K2CO3 (DMF-d7, 25 °C) provided a 3-5:95-97 mixture of
(P)-30:(P)-24 in the time that it takes to immediately record the 1H
NMR. Conformational searches conducted on both atropisomers of the
8R and 8S diastereomers 30/24, 28/29, 37/45 (MacroModel BatchMin
6.0, OPLSA Force Field) revealed a preference for the unnatural 8S
diastereomer (∆E ) 2.3-3.9 kcal/mol) which is consistent with the
experimental observations, the epimerization of 30 to 24, and the
assigned stereochemistries. The natural 8R-37 proved more resistant
toward C8 epimerization than 30. Thus, treatment of 37 with 2 equiv
of DBU (THF, 25 °C, 40 min), 5 equiv of K2CO3 (DMF, 45 °C, 12 h), or
5 equiv CsF (DMF, 25 °, 8 h) led to generation of 20%, 58%, and 5%,
respectively, of the C8 diastereomer 45 and the equilibrium ratio of
8S:8R was established to be 95:5. Data for tert-butyl (M)- and (P)-(8R,-
11S,14R,15R)-14-[N-[(tert-butyloxy)carbonyl]amino]-11-(cyanomethyl)-
5, 15,-dihydroxy-10,13-dioxo-4-methoxy-18-nitro-9,12-diaza-2-oxatricyclo-
[14.2.2.13,7]heneicosa-3,4,7(21),16,18,19-hexaene-8-carboxylate (30) follow.
(M)-30 (the less polar isomer): [R]25 +78 (c 0.22, CH3OH); 1H NMR
D
(acetone-d6, 400 MHz) δ 8.28-8.21 (m, 1H), 8.20 (s, 1H), 8.16 (s, 1H),
7.96 (d, 1H, J ) 8.1 Hz), 7.53-7.47 (m, 1H), 7.44 (d, 1H, J ) 8.4 Hz),
6.60 (s, 1H), 6.27 (d, 1H, J ) 8.4 Hz), 5.46 (d, 1H, J ) 1.9 Hz), 5.27 (s,
1H), 5.14-5.06 (m, 1H), 4.99 (d, 1H, J ) 7.0 Hz), 4.73-4.68 (m, 1H),
4.62-4.55 (m, 1H), 3.95 (s, 3H), 2.93-2.77 (m, 2H), 1.44 (s, 9H), 1.40
(s, 9H); IR (neat) νmax 3326, 1660, 1537, 1341, 1249, 1158 cm-1
;
FABHRMS (NBA-CsI) m/z 804.1529 (M+ + Cs, C31H37N5O12 requires
804.1493). The 2D 1H-1H ROESY NMR spectrum (CD3OD, 600 MHz,
25 °C) of (M)-30 exhibited the following diagnostic NOE cross-peaks:
H-20/H-19 (s, δ 7.75/7.35), H-20/H-15 (s, δ 7.75/5.15), H-20/H-14 (s, δ
7.75/4.56), H-15/H-14 (s, δ 5.15/4.56), H-6/H-8 (m, δ 6.50/4.77), H-11/
CH2CN (m, δ 4.48/2.84 and 4.48/2.71). For (P)-30 (the more polar
(21) The chemical shifts of C6-H, C21-H, C8-H, and N9-H shift
dramatically (>1 ppm) and the signals previously attributable to C8-H
and N9-H collapse to singlets versus doublets. For the major atropi-
somer derived from 15: [R]25D -11.6 (c 0.01, CHCl3); 1H NMR (acetone-
d6, 400 MHz) δ 9.22 (s, 1H), 8.21 (d, 1H, J ) 8.2 Hz), 8.11 (s, 1H), 7.76
(d, 1H, J ) 8.7, 2.2 Hz), 7.36 (d, 1H, J ) 2.1 Hz), 7.26 (br s, 1H), 7.13
(s, 1H), 7.11 (d, 1H, J ) 7.2 Hz), 6.76 (br s, 1H), 6.47 (d, 1H, J ) 8.3
Hz), 5.47 (d, 1H, J ) 4.2 Hz), 5.11-5.09 (m, 1H), 4.77-4.71 (m, 1H),
4.35 (t, 1H, J ) 8.5 Hz), 3.92 (s, 3H), 3.05 (dd, 1H, J ) 17.1, 5.3 Hz),
2.93 (dd, 1H, J ) 17.1, 7.6 Hz), 1.56 (s, 9H), 1.30 (s, 9H); IR (film) νmax
3313, 2923, 1676, 1539, 1349, 1246, 1133 cm-1; FABMS (NBA-NaI)
m/z 710 (C31H37N5O12 requires 694). For the product derived from 20
(40%): 1H NMR (acetone-d6, 400 MHz) δ 9.09 (s, 1H), 8.18-8.14 (m,
2H), 7.97 (d, 1H, J ) 8.0 Hz), 7.77 (dd, 1H, J ) 8.8, 2.1 Hz), 7.35 (d,
1H, J ) 2.0 Hz), 7.28 (s, 1H), 7.19-7.16 (m, 4H), 7.12-7.08 (m, 2H),
6.88-6.82 (m, 4H), 6.55 (s, 1H), 6.22 (d, 1H, J ) 9.2 Hz), 6.10 (d, 1H,
J ) 8.3 Hz), 5.80 (d, 1H, J ) 3.5 Hz), 4.94-4.90 (m, 1H), 4.78-4.74
(m, 1H), 4.33 (t, J ) 8.6 Hz), 3.86 (s, 3H), 3.76 (s, 3H), 3.75 (s, 3H),
1.54 (s, 9H), 1.23 (s, 9H); FABMS (NBA-NaI) m/z 955 (M+ + Na,
isomer): [R]25 +65 (c 0.18, CH3OH); 1H NMR (acetone-d6, 400 MHz)
D
δ 8.25 (s, 1H), 8.20 (s, 1H), 7.94 (d, 1H, J ) 5.6 Hz), 7.82 (d, 1H, J )
7.3 Hz), 7.76 (d, 1H, J ) 8.4 Hz), 7.26 (d, 1H, J ) 8.6 Hz), 6.66 (s, 1H),
6.34 (d, 1H, J ) 7.6 Hz), 5.69 (s, 1H), 5.31-5.26 (m, 1H), 5.13 (d, 1H,
J ) 7.8 Hz), 5.07 (d, 1H, J ) 6.8 Hz), 4.68-4.60 (m, 1H), 4.56 (d, 1H,
J ) 5.6 Hz), 3.94 (s, 3H), 2.85-2.75 (m, 2H), 1.44 (s, 9H), 1.38 (s, 9H);
IR (neat) νmax 3410, 3287, 1668, 1634, 1537, 1254, 1158 cm-1
;
FABHRMS (NBA-CsI) m/z 804.1529 (M+ + Cs, C31H37N5O12 requires
804.1493). The 2D 1H-1H ROESY NMR spectrum (CD3OD, 600 MHz,
25 °C) of (P)-30 exhibited the following diagnostic NOE cross-peaks:
H-17/H-15 (s, δ 8.12/5.14), H-17/H-14 (s, δ 8.12/4.41), H-20/H-19 (s, δ
7.79/7.17), H-19/H-21 (m, δ 7.17/5.62), H-15/H-14 (s, δ 5.14/4.41), H-6/
H-8 (s, δ 6.53/4.79), H-11/CH2CN (m, δ 4.52/2.78 and 4.52/2.70).
(20) Tentative C8 stereochemical assignments for 16 and 21 are also
the unnatural, epimerized 8S stereochemistry. A conformational search
of both atropisomers of the C8 diastereomers of 16 (MacroModel
BatchMin 6.0, OPLSA) established that the unnatural 8S diastereo-
mers were substantially more stable (∆E ) >2.5 kcal/mol). For 21,
this is analogous to observations made with 4 and further supported
by its relative ease of atropisomerism, see Table 2. Product 14 has
been tentatively assigned the natural 8R stereochemistry on the basis
of a conformational search (MacroModel, BatchMin 6.0, OPLSA)
conducted on both atropisomers of the C8 diastereomers in which the
natural, nonepimerized 8R configuration was established to be more
stable (∆E ) 1.1-2.7 kcal/mol).
C
46H53O15N5 requires 939). For the products derived from 31 (70%):
1H NMR (acetone-d6, 400 MHz) δ 9.09 (s, 1H), 8.19 (d, 1H, J ) 7.8
Hz), 8.14 (s, 1H), 7.77 (dd, 1H, J ) 8.5, 2.2 Hz), 7.35 (d, 1H, J ) 2.0
Hz), 7.35-7.33 (m, 1H), 7.12 (s, 1H), 7.11 (d, 1H, J ) 8.6 Hz), 6.78-
6.76 (m, 1H), 6.31 (d, 1H, J ) 9.0 Hz), 5.61 (d, 1H, J ) 3.1 Hz), 5.01
(d, 1H, J ) 9.0 Hz), 4.80-4.76 (m, 1H), 4.43 (t, 1H, J ) 9.0 Hz), 3.92
(s, 3H), 2.98-2.82 (m, 2H), 1.56 (s, 9H), 1.18 (s, 9H); IR (film) νmax
3313, 2923, 2246, 1676, 1539, 1349, 1133 cm-1; FABMS (NBA-NaI)
m/z 710 (M+ + Na, C31H37O12N5 requires 694).
(22) Ko¨nig, W.; Geiger, R. Chem. Ber. 1970, 103, 2041. Prepared by
treatment of CBZNH-Asn with bis(4-methoxyphenyl)carbinol (cat. H2-
SO4, HOAc, 25 °C, 13.5 h, 84%).