Second-generation DBFOX ligands for the synthesis of β-substituted α-amino acids via enantioselective radical conjugate additions

Article abstract of DOI:10.1021/jo801721z

(Chemical Equation Presented) A set of second-generation DBFOX ligands possessing extended aryl or benzyl-type groups was synthesized. The requisite amino alcohols were either commercially available (DBFOX/Bn) or constructed via Sharpless asymmetric amino

Full text of DOI:10.1021/jo801721z

Second-Generation DBFOX Ligands for the Synthesis of
ꢀ-Substituted r-Amino Acids via Enantioselective Radical Conjugate
Additions
Biplab Banerjee, Steven G. Capps, Junghoon Kang, Joshua W. Robinson, and
Steven L. Castle*
Department of Chemistry and Biochemistry, Brigham Young UniVersity, ProVo, Utah 84602
scastle@chem.byu.edu
ReceiVed August 1, 2008
A set of second-generation DBFOX ligands possessing extended aryl or benzyl-type groups was
synthesized. The requisite amino alcohols were either commercially available (DBFOX/Bn) or constructed
via Sharpless asymmetric aminohydroxylation (DBFOX/Nap, DBFOX/t-BuPh, DBFOX/Pip) or phase-
transfer-catalyzed asymmetric alkylation (DBFOX/MeNap). Complexes of the ligands with Mg(NTf2)2
were evaluated as promoters of enantioselective radical conjugate additions to R,ꢀ-unsaturated R-nitro
amides and esters. Reactions employing the DBFOX/Nap ligand exhibited improved enantioselectivity
relative to previously published additions mediated by DBFOX/Ph. However, the relatively modest increase
in diastereomeric ratio suggests that our substrate-Lewis acid binding model, which was formulated
based on results from DBFOX/Ph-promoted radical conjugate additions, is in need of revision.
Introduction
conditions with acidic amide hydrogens and other functional
groups present in peptides.7 In theory, this feature would allow
generation of a ꢀ-substituted R-amino acid from a complex
peptidic substrate possessing the requisite radical acceptor.
Consequently, this method should have great utility in the total
synthesis of peptide natural products.
Recently, we discovered that the DBFOX/Ph ligand developed
by Kanemasa and Curran (1, Figure 1)1 is uniquely effective at
facilitating the synthesis of ꢀ-substituted R-amino acids via
enantioselective Lewis acid promoted radical conjugate additions
(Scheme 1).2 ꢀ-Substituted R-amino acids are constituents of
several peptide natural products;3 additionally, they are valued
as conformationally constrained analogues of R-amino acids.4
Therefore, numerous means have been devised for their con-
struction.5 The radical conjugate addition approach6 to these
compounds is attractive due to compatibility of the mild reaction
The primary drawback to the DBFOX/Ph-promoted radical
conjugate additions is the lack of diastereoselectivity. Although
the adducts could be obtained in reasonable ee values, the
diastereomeric ratios were uniformly poor (<2:1 in almost every
case).2 After determining the absolute configurations of the
adducts, we formulated an empirical substrate-Lewis acid
binding model (Figure 2). This model is consistent with Curran
and Kanemasa’s proposal of octahedral geometry for Mg-
(ClO4)2-DBFOX/Ph complexes.1b Additionally, it accounts for
the fact that the hydrogen atom abstraction step occurring at
the substrate R-carbon is much more stereoselective than the
addition step, which takes place at the ꢀ-carbon.2 Apparently,
the ligand is shielding the R-carbon more efficiently than the
ꢀ-carbon. We reasoned that modifying the ligand by extending
the size of the phenyl groups would lead to increased diaster-
(1) (a) Kanemasa, S.; Oderaotoshi, Y.; Yamamoto, H.; Tanaka, J.; Wada,
E.; Curran, D. P. J. Org. Chem. 1997, 62, 6454. (b) Iserloh, U.; Curran, D. P.;
Kanemasa, S. Tetrahedron: Asymmetry 1999, 10, 2417. (c) Iserloh, U.;
Oderaotoshi, Y.; Kanemasa, S.; Curran, D. P. Org. Synth. 2003, 80, 46.
(2) He, L.; Srikanth, G. S. C.; Castle, S. L. J. Org. Chem. 2005, 70, 8140.
(3) (a) Suzuki, H.; Morita, H.; Shiro, M.; Kobayashi, J. Tetrahedron 2004,
60, 2489. (b) Suzuki, H.; Morita, H.; Iwasaki, S.; Kobayashi, J. Tetrahedron
2003, 59, 5307. (c) Kobayashi, J.; Suzuki, H.; Shimbo, K.; Takeya, K.; Morita,
H. J. Org. Chem. 2001, 66, 6626. (d) Leung, T.-W. C.; Williams, D. H.; Barna,
J. C. J.; Foti, S.; Oelrichs, P. B. Tetrahedron 1986, 42, 3333.
(4) (a) Hruby, V. J. J. Med. Chem. 2003, 46, 4215. (b) Gibson, S. E.; Guillo,
N.; Tozer, M. J. Tetrahedron 1999, 55, 585.
10.1021/jo801721z CCC: $40.75
Published on Web 10/24/2008
2008 American Chemical Society
J. Org. Chem. 2008, 73, 8973–8978 8973

Banerjee et al.
SCHEME 2. Synthesis of Aryl-Type DBFOX Ligands
FIGURE 1. DBFOX/Ph.
SCHEME 1. Mg/1-Promoted Enantioselective Radical
Conjugate Additions
FIGURE 2. Substrate-Lewis acid binding model.
eomeric ratios due to enhanced shielding of the substrate
ꢀ-carbon. Herein, we present the synthesis of second-generation
DBFOX ligands bearing extended aryl or benzyl groups and
an evaluation of their performance in the radical conjugate
addition reaction.
The final step in the Curran-Kanemasa synthesis of 1 is a
high-yielding (94%) bis-cyclodehydration mediated by diethy-
laminosulfur trifluoride (DAST).1c However, DAST-mediated
cyclodehydrations of diamides 5a-c were capricious and low-
yielding. It is reported that recrystallization of the diamide
precursor to 1 is essential for obtaining good yields of the
DBFOX/Ph ligand by this method.1c Recrystallizations of 5a-c
were impractical due to the small scale of this exploratory
project; consequently, the diamides were purified by silica gel
chromatography. Thus, it is likely that trace impurities present
in our samples of 5a-c are responsible for the poor results with
DAST as the cyclodehydration agent. Fortunately, treatment
of the diamides with Et3N, TsCl, and a catalytic amount of
4-pyrrolidinopyridine (PPY)9 allowed reliable production of
Results and Discussion
The initial set of ligands that we targeted replaced the phenyl
groups of 1 with larger aryl groups. The synthesis of 1 reported
by Curran and Kanemasa is quite straightforward, involving bis-
amidation of dibenzofuran-4,6-dicarbonyl chloride (4, Scheme
2) with (R)-phenylglycinol followed by bis-cyclodehydration
to form the oxazoline rings.1c Thus, to replace the phenyl groups
of 1, we required enantiopure arylglycinols other than phenylg-
lycinol. We elected to synthesize these intermediates via
Sharpless asymmetric aminohydroxylation (SAA)8 of the cor-
responding styrenes, as shown in Scheme 2. The SAA of
2-vinylnaphthalene (2a) has been performed previously,8b and
afforded 3a in good yield and excellent ee. Use of the 4-tert-
butylphenyl (2b) and piperonyl (2c) congeners resulted in lower
yields, but the ee values remained excellent. In each case, the
desired N-Cbz amino alcohol 3a-c could be isolated by
crystallization from the reaction mixture. In some instances the
regioisomeric amino alcohol could be detected as a minor
product in the reaction mixture, but its isolation and quantifica-
tion was not pursued.
The Cbz groups of 3a-c were cleaved via hydrogenolysis,
and the resulting crude amino alcohols were subjected to
amidation with 4. The amidation of 4 with (R)-naphthylglycinol
derived from 3a proceeded in analogous fashion to the reported
reaction with (R)-phenylglycinol.1c However, amidations em-
ploying the amino alcohols derived from 3b and 3c were
sluggish, requiring elevated temperatures and extra time for
completion. Presumably, the increased steric bulk of these amino
alcohols relative to phenylglycinol is responsible for the
decreased reactivity, although it is unclear to us why naphth-
ylglycinol behaves similarly to phenylglycinol.
(5) (a) Qu, H.; Gu, X.; Liu, Z.; Min, B. J.; Hruby, V. J. Org. Lett. 2007, 9,
3997. (b) Ooi, T.; Kato, D.; Inamura, K.; Ohmatsu, K.; Maruoka, K. Org. Lett.
2007, 9, 3945. (c) Bull, S. D.; Davies, S. G.; Epstein, S. W.; Garner, A. C.;
Mujtaba, N.; Roberts, P. M.; Savory, E. D.; Smith, A. D.; Tamayo, J. A.; Watkin,
D. J. Tetrahedron 2006, 62, 7911. (d) Gu, X.; Ndungu, J. M.; Qiu, W.; Ying, J.;
Carducci, M. D.; Wooden, H.; Hruby, V. J. Tetrahedron 2004, 60, 8233. (e)
Spangenberg, T.; Schoenfelder, A.; Breit, B.; Mann, A. Org. Lett. 2007, 9, 3881.
(f) Yu, S.; Pan, X.; Ma, D. Chem. Eur. J. 2006, 12, 6572. (g) Roff, G. J.; Lloyd,
R. C.; Turner, N. J. J. Am. Chem. Soc. 2004, 126, 4098. (h) Burk, M. J.;
Bedingfield, K. M.; Kiesman, W. F.; Allen, J. G. Tetrahedron Lett. 1999, 40,
3093. (i) Jamieson, A. G.; Sutherland, A.; Willis, C. L. Org. Biomol. Chem.
2004, 2, 808. (j) O’Donnell, M. J.; Cooper, J. T.; Mader, M. M. J. Am. Chem.
Soc. 2003, 125, 2370. (k) Zhang, H.; Mitsumori, S.; Utsumi, N.; Imai, M.; Garcia-
Delgado, N.; Mifsud, M.; Albertshofer, K.; Cheong, P. H.-Y.; Houk, K. N.;
Tanaka, F.; Barbas, C. F., III. J. Am. Chem. Soc. 2008, 130, 875. (l) Wang, W.;
Wang, J.; Li, H. Tetrahedron Lett. 2004, 45, 7243. (m) Notz, W.; Tanaka, F.;
Watanabe, S.; Chowdari, N. S.; Turner, J. M.; Thayumanavan, R.; Barbas, C. F.,
III. J. Org. Chem. 2003, 68, 9624. (n) Marigo, M.; Kjærsgaard, A.; Juhl, K.;
Gathergood, N.; Jørgensen, K. A. Chem. Eur. J. 2003, 9, 2359. (o) Ferraris, D.;
Young, B.; Cox, C.; Dudding, T.; Drury, W. J., III.; Ryzhkov, L.; Taggi, A. E.;
Lectka, T. J. Am. Chem. Soc. 2002, 124, 67. (p) Hara, S.; Makino, K.; Hamada,
Y. Tetrahedron 2004, 60, 8031. (q) Acevedo, C. M.; Kogut, E. F.; Lipton, M. A.
Tetrahedron 2001, 57, 6353. (r) Han, G.; Lewis, A.; Hruby, V. J. Tetrahedron
Lett. 2001, 42, 4601. (s) Liang, B.; Carroll, P. J.; Joullie´, M. M. Org. Lett. 2000,
2, 4157. (t) Burtin, G.; Corringer, P.-J.; Young, D. W. J. Chem. Soc., Perkin
Trans. 1 2000, 3451. (u) Tamura, O.; Yoshida, S.; Sugita, H.; Mita, N.; Uyama,
Y.; Morita, N.; Ishiguro, M.; Kawasaki, T.; Ishibashi, H.; Sakamoto, M. Synlett
2000, 1553. (v) Medina, E.; Moyano, A.; Perica`s, M. A.; Riera, A. HelV. Chim.
Acta 2000, 83, 972.
8974 J. Org. Chem. Vol. 73, No. 22, 2008

Synthesis of ꢀ-Substituted R-Amino Acids
TABLE 1. Evaluation of Ligands in Radical Conjugate Additiona
SCHEME 3. Synthesis of DBFOX/Bn
ligand
yield (%)
syn/anti
% ee (syn, anti)b
1c
6a
6b
6c
9
76
65
57
75
80
44
1.4:1
1.8:1
1.6:1
1.2:1
1.5:1
1.6:1
88, 76
97, 92
84, 79
76, 79
82, 80
92, 90d
13
a 1.0 equiv of Mg(NTf2)2 and ligand were employed in each reaction.
b Determined by chiral HPLC (see the Experimental Section for details).
c Data from ref 2. d Major enantiomers were opposite those obtained
from reactions employing 1, 6a-c, and 9.
SCHEME 4. Synthesis of DBFOX/MeNap
higher temperatures and longer reaction times to yield a
satisfactory amount of amide 12. The cyclization of 12 to form
13 proceeded under the conditions employed for DBFOX/Bn
(9). Ligand 13 was prepared as the opposite enantiomer relative
to 6a-c and 9 due to the ready availability of the cinchonidine-
derived phase-transfer catalyst (as opposed to the cinchonine-
derived catalyst) in our laboratory.
With ligands 6a-c, 9, and 13 in hand, we turned our attention
to evaluating their performance in the enantioselective conjugate
addition of isopropyl radical to R,ꢀ-unsaturated R-nitroamide
14a (Table 1). To facilitate analysis, the crude adduct was
reduced to the corresponding amine and protected, affording
carbamate 15a as a mixture of diastereomers. The data from
this investigation are presented in Table 1 alongside our
previously reported results with DBFOX/Ph (1).2 The DBFOX/
Nap ligand (6a) afforded both diastereomers of 15a in excellent
ee (97% for syn-15a, 92% for anti-15a) and acceptable yield.
DBFOX/t-BuPh (6b), DBFOX/Pip (6c), and DBFOX/Bn (9)
each provided the adducts in comparable or lower ee than did
DBFOX/Ph. DBFOX/MeNap (13) delivered both diastereomers
of 15a in >90% ee; however, the combined yield was low
(44%). This reduction in yield is likely a consequence of the
increased bulk of ligand 13. The reactions listed in Table 1 were
conducted in identical fashion to previously reported radical
conjugate additions employing ligand 12 with the exception that
longer times were required to ensure complete complexation
of Mg(NTf2)2 to the new, bulkier ligands. Conversion of the
Mg(NTf2)2/ligand/CH2Cl2 mixture from a cloudy suspension to
a clear solution was used as evidence that complexation had
been achieved. Reactions initiated prior to observation of
complexation afforded products derived from reduction of the
double bond rather than from radical conjugate addition.
Presumably, Mg(NTf2)2 is a stronger Lewis acid than the Mg/
DBFOX complex, which possesses a large, relatively electron-
rich ligand. Accordingly, the stronger Lewis acid (Mg(NTf2)2)
DBFOX/Nap (6a), DBFOX/t-BuPh (6b), and DBFOX/Pip (6c).
Notably, the inclusion of PPY was critical, as use of the weaker
nucleophilic catalyst DMAP resulted in sluggish, low-yielding
reactions.
Later, we decided to expand our study to include DBFOX
ligands in which the aryl group is separated from the oxazoline
ring by a methylene spacer. Accordingly, commercially available
(R)-phenylalaninol (7, Scheme 3) was subjected to the two-
step amidation-cyclodehydration protocol, yielding DBFOX/
Bn (9). The synthesis of 9 proceeded in similar fashion to the
preparation of 6b and 6c with the exception that additional
quantities of PPY (20 mol %) were required for the cyclode-
hydration. DBFOX/MeNap (13), which incorporates a naphthyl
group attached to the methylene spacer, was constructed as
outlined in Scheme 4. The requisite naphthylalanine derivative
10 was available in high ee via a previously reported enanti-
oselective alkylation mediated by a chiral phase-transfer cata-
lyst.10 Hydrolysis of the benzophenone imine followed by
reduction of the tert-butyl ester provided amino alcohol 11.11
The bis-amidation of 4 with 11 was more sluggish than the
analogous reactions used to synthesize 5a-c and 8, requiring
(7) For syntheses of racemic ꢀ-substituted R-amino acids via radical conjugate
additions, see: (a) Renaud, P.; Stojanovic, A. Tetrahedron Lett. 1996, 37, 2569.
(b) Srikanth, G. S. C.; Castle, S. L. Org. Lett. 2004, 6, 449.
(8) (a) Reddy, K. L.; Sharpless, K. B. J. Am. Chem. Soc. 1998, 120, 1207.
(b) Li, G.; Lenington, R.; Willis, S.; Kim, S. H. J. Chem. Soc., Perkin Trans. 1
1998, 1753.
(9) Evans, D. A.; Woerpel, K. A.; Nosse, B.; Schall, A.; Shinde, Y.; Jezek,
E.; Haque, M. M.; Chhor, R. B.; Reiser, O. Org. Synth. 2006, 83, 97.
(10) (a) Tomotaka, O.; Takemoto, Y. Org. Lett. 2001, 3, 1515. (b) Lee, J.-
H.; Yoo, M.-S.; Jung, J.-H.; Jew, S.-S.; Park, H.-G. Tetrahedron 2007, 63, 7906.
(11) Combret, Y.; Duflos, J.; Dupas, G.; Bourguignon, J.; Que´guiner, G.
Tetrahedron: Asymmetry 1993, 4, 1635.
(6) For a review, see: Srikanth, G. S. C.; Castle, S. L. Tetrahedron 2005,
61, 10377.
J. Org. Chem. Vol. 73, No. 22, 2008 8975

Banerjee et al.
TABLE 2. Scope of Additions with DBFOX/Nap
TABLE 3. Selectivity of r and ꢀ Stereocenter Formation in
Reactions with Ligand 6aa
substrate
R (R):(S)
ꢀ (R):(S)
14a
14b
14c
98:2 (92:8)
89:11 (82:18)
84:16 (82:18)
35:65 (40:60)
43:57 (44:56)
37:63 (37:63)
a Values in parentheses are data from reactions with ligand 1 (ref 2).
extend this assignment by analogy to the adducts derived from
substrates 14a and 14c (a reasonable assumption given the
similarities in chiral HPLC elution profiles), we can calculate
the selectivity of the radical addition and hydrogen atom ab-
straction steps. These values are expressed in ratios as “ꢀ (R):
(S)” and “R (R):(S)”, respectively, in Table 3. The binding model
predicted that extension of the ligand aryl groups would lead
to higher syn/anti ratios due to a more selective addition process.
However, the selectivity of this step increased very slightly; in
fact, improvements in the hydrogen atom abstraction were more
significant, albeit still modest. The fact that the high syn/anti
ratios forecast by the binding model failed to materialize
suggests that the substrates are interacting with the chiral Lewis
acid in a manner different from that depicted in Figure 2.
Moreover, the dramatic contrast in additions to amides 14
versus esters 16 may indicate that the ester substrates bind
only weakly to the chiral Lewis acid, thereby allowing the
reaction to proceed at least partially through a nonselective
pathway with uncomplexed substrates. Indeed, our prior work
demonstrated that a relatively slow radical conjugate addition
can occur in the absence of Lewis acid promoters.2 R,ꢀ-
Unsaturated amides have been shown to complex to Lewis
acids more strongly than the corresponding esters.12 This
lower affinity of esters versus amides for Lewis acids could
lead to slower, less-selective reactions with substrates 16.
Additionally, the lower yields and ee values of additions to
16a-c with 6a in place of 1 could be a consequence of the
extra bulk of the DBFOX/Nap ligand further discouraging
ester-Lewis acid complexation. These data suggest that the
major determining factor in the stereoselective radical
conjugate additions is the strength of Lewis acid binding to
the carbonyl group. In contrast, the nitro group may be
interacting weakly or not at all with the chiral Lewis acid.
Although complexation of nitro groups to Lewis acids has
been reported previously,13 the minimal improvement in ꢀ
(R):(S) with ligand 6a casts doubt on the existence of
bidentate chelation involving the nitro moiety as portrayed
in Figure 2.
substrate
ligand
yield (%)
syn/anti
% ee (syn, anti)a
14a
14a
14b
14b
14c
14c
16a
16a
16b
16b
16c
16c
6a
1b
6a
1b
6a
1b
6a
1b
6a
1b
6a
1b
65
76
60
66
63
59
38
75
28
60
17
66
1.8:1
1.4:1
1.7:1
1.4:1
2.0:1
1.6:1
2.2:1
2.6:1
1.3:1
1.7:1
1.4:1
1.7:1
97, 92
88, 76
74, 87
64, 62
70, 63
72, 50
10, 7
21, 28
7, 5
9, 17
3, 3
6, 12
a Determined by chiral HPLC (see the Experimental Section for
details). b Data from ref 2.
promotes conjugate reduction of 14a by Bu3SnH, whereas the
weaker Lewis acid (Mg/DBFOX) mediates the desired radical
conjugate addition. In fact, conjugate reduction of 14a and
related acceptors was observed in our prior work when strong
Lewis acids such as MgBr2 ·OEt2 and lanthanide triflates were
employed.2
We were disappointed by the lack of significant improvement
in reaction diastereoselectivity (1.8:1 dr with 6a compared to
1.4:1 dr with 1). Nevertheless, the increased ee values obtained
with DBFOX/Nap warranted further investigations of enanti-
oselective radical conjugate additions with this ligand. Conse-
quently, we examined Mg-DBFOX/Nap-mediated additions of
isopropyl radical to acceptors other than 14a. The results of
these reactions are summarized in Table 2 along with data from
the analogous reactions employing ligand 1.2 Amide substrates
14b,c underwent Mg/6a-promoted radical conjugate additions
to provide adducts 15b,c in equivalent or greater ee than was
observed in Mg/1-mediated reactions. Additionally, the syn/anti
ratios were slightly higher with the new ligand. Both ligands
gave best results with substrate 14a bearing an electron-rich
p-methoxyphenyl group at the alkene ꢀ-position. Unfortunately,
Mg/6a-promoted additions to ester substrates 16a-c were
characterized by significantly lower yields and ee values than
resulted from the previous reactions employing Mg/1 as the
chiral Lewis acid.
Chiral HPLC elution profiles established that the same major
enantiomers are produced from reactions utilizing either DB-
FOX/Ph or DBFOX/Nap. Thus, it appears the chiral Lewis acids
formed by complexation of these ligands with Mg(NTf2)2 are
binding to the substrates in the same fashion. However, our new
results do indicate that the binding model shown in Figure 2
may need revision. We previously determined the absolute
configuration of the adducts derived from substrate 14b.2 If we
For substrates 14 and 16 to coordinate to the Mg-
(NTf2)2-DBFOX complex in a bidentate fashion, one of the
triflimide ligands must be displaced from the Lewis acid (Figure
3). In an attempt to probe the structure of the complex, we
conducted 19F NMR experiments in CD2Cl2.14 The 19F NMR
spectrum of the Mg(NTf2)2-DBFOX/Nap complex consisted
of a single peak at temperatures ranging from 20 to -90 °C.
(12) Urabe, H.; Yamashita, K.; Suzuki, K.; Kobayashi, K.; Sato, F. J. Org.
Chem. 1995, 60, 3576.
(13) (a) Liu, H.; Xu, J.; Du, D.-M. Org. Lett. 2007, 9, 4725. (b) Lu, S.-F.;
Du, D.-M.; Xu, J. Org. Lett. 2006, 8, 2115. (c) Singh, P. K.; Bisai, A.; Singh,
V. K. Tetrahedron Lett. 2007, 48, 1127. (d) Jia, Y.-X.; Zhu, S.-F.; Yang, Y.;
Zhou, Q.-L. J. Org. Chem. 2006, 71, 75. (e) Wada, E.; Yoshinaga, M.
Tetrahedron Lett. 2004, 45, 2197. (f) Wada, E.; Yoshinaga, M. Tetrahedron
Lett. 2003, 44, 7953.
(14) For an NMR study of Zn-DBFOX complexes with 3-acetyl-2-oxazoli-
dinone, see: Kanemasa, S.; Oderaotoshi, Y.; Tanaka, J.; Wada, E. Tetrahedron
Lett. 1998, 39, 7521.
8976 J. Org. Chem. Vol. 73, No. 22, 2008

Synthesis of ꢀ-Substituted R-Amino Acids
was made to optimize other parameters such as reaction time;
thus, it is conceivable that the yield could be restored to its
previous level or even increased by careful experimentation.
Although further work is necessary, it appears that the radical
conjugate addition can be rendered more practical by decreasing
the reagent amounts.
Conclusions
FIGURE 3. Potential substrate-Lewis acid binding modes.
Prompted by an empirical substrate-Lewis acid binding
model developed to explain the results of our Mg-DBFOX/
Ph-promoted enantioselective radical conjugate additions, we
have designed and synthesized second-generation DBFOX
ligands bearing extended aryl or benzyl groups. Although
the DBFOX/Nap ligand (6a) did afford improved enantiose-
lectivity in the radical conjugate addition, the increase in
diastereomeric ratio was disappointingly small. Accordingly,
it seems that the substrate-Lewis acid binding model is in
need of revision. The available data from reactions and NMR
experiments suggest that the nitro group of substrates 14 and
16 is not complexing strongly to the Lewis acid. Despite the
fact that these ligands did not perform as anticipated in the
radical conjugate addition, they may be useful in other
asymmetric reactions. In particular, the concise syntheses of
6a-c, 9, and 13 reported herein (2-4 steps from com-
mercially available or known materials) makes them easily
accessible. Consequently, the utility of these second-genera-
tion DBFOX ligands in other transformations as well as the
design of new promoters for the radical conjugate addition
will be subjects of future investigations.
SCHEME 5. Reduced Tin Loadings
Then, amide 14a was added to the solution, which was stirred
at -78 °C for 30 min to promote complexation. These are the
identical conditions used to form the complex prior to initiation
of the radical conjugate addition. Again, 19F NMR spectra of
the resulting complex exhibited only one signal across the
aforementioned temperature range. These observations are
consistent with a monodentate binding mode of substrate 14a,
which does not require displacement of a triflimide ligand. If
the substrate were bound in a bidentate fashion, then two 19F
signals should have been observed due to the presence of both
bound and free triflimide ions. We cannot rule out the possibility
of rapid triflimide ligand exchange on the complex; however,
our data require any such exchange to be rapid even at -90
°C. The precise structure of a monodentate substrate-Lewis
acid complex would be hard to characterize, as numerous modes
of binding can be envisioned. Nonetheless, it is plausible that
the existence of either a relatively loose monodentate complex
or multiple types of monodentate complexes would have a
negative impact on the stereoselectivity of the radical conjugate
addition.
It should be noted that R,ꢀ-unsaturated amides 14a-c were
employed as the E-isomers, whereas the Z-isomers of esters
16a-c were utilized in the radical conjugate additions. The
amides were obtained exclusively as E-isomers from Knoev-
enagel-type condensations, while the esters were produced as
ca. 2:1 mixtures of olefinic isomers from which Z-16a-c could
be isolated via recrystallization.2 Previously conducted additions
to both isomerically pure and impure samples of 16a-c
employing achiral Lewis acids provided the adducts in 1:1 dr.2
Thus, it appears that the newly formed ꢀ-stereocenter does not
impact the subsequent hydrogen atom abstraction. This phe-
nomenon could be attributed to the fact that the aryl and
isopropyl groups attached to the ꢀ-carbon are of roughly similar
size. Although the olefin geometry does not impact the diaster-
eomeric ratio, it is possible that the E-amides 14 may bind to
the chiral Lewis acid in a different manner than do the Z-esters
16.
Experimental Section
(R)-Benzyl 2-Hydroxy-1-(naphthalen-2-yl)ethylcarbamate (3a).
A stirred solution of benzyl carbamate (309.0 mg, 2.04 mmol) in
n-propanol (2 mL) was treated with a freshly prepared NaOH
solution (1.25 M in H2O, 1.0 mL, 1.25 mmol). tert-Butyl hypochlo-
rite15 (240.7 mg, 2.22 mmol) was then added, followed by
(DHQD)2PHAL (30.9 mg, 0.040 mmol). The resulting mixture was
stirred until homogeneous, then cooled to 0 °C and stirred for 10
min. 2-vinylnaphthalene (206.0 mg, 1.34 mmol) was then added,
followed by a solution of K2OsO2(OH)4 (10.0 mg, 0.027 mmol) in
1.25 M NaOH (1.0 mL, 1.25 mmol). The resulting solution was
stirred at 0 °C for 4.5 h, at which time stirring was ceased and the
flask was cooled to -25 °C. The product was collected by filtration,
washed with cold n-PrOH-H2O (1:1), and dried overnight on the
benchtop to afford 3a (288.3 mg, 0.90 mmol, 68%) as a white solid.
Spectral data for 3a were identical with those previously reported.8b,16
N4,N6-Bis((R)-2-hydroxy-1-(naphthalen-2-yl)ethyl)dibenzo-
[b,d]furan-4,6-dicarboxamide (5a). A solution of 3a (161.0 mg, 0.50
mmol) in MeOH (4.0 mL) was treated with NH4OAc (117.9 mg,
1.53 mmol) followed by 10% Pd/C (21.4 mg, 0.13 wt equiv). The
resulting mixture was stirred at rt under H2 (1 atm) for 21 h, then
(15) Mintz, M. J.; Walling, C. Organic Syntheses; Wiley: New York, 1973;
Collect. Vol. V, p 184. For best results, this reagent was used within two weeks
of preparation.
The reactions summarized in Tables 1 and 2 were conducted
with substantial quantities of Bu3SnH (three portions of 2.5
equiv each) to ensure complete conversion. With the goal of
enhancing the preparative value of this transformation, we
performed the DBFOX/Nap-mediated addition of isopropyl
radical to 14a with reduced tin loadings (Scheme 5). The dr
and ee values were comparable to those obtained from the earlier
experiments, but the yield diminished somewhat. No attempt
(16) Occasionally, the product would fail to crystallize from the reaction
mixture. In these cases, the following workup and isolation procedure was
employed. The solution was treated with saturated aq Na2SO3 (10 mL), then
stirred at 0 °C for 15 min. The layers were separated, and the aqueous layer was
extracted with EtOAc (3 × 15 mL). The combined organic layers were washed
with H2O (10 mL) and brine (10 mL), dried (MgSO4), and concentrated in vacuo.
Flash chromatography (SiO2, 2 × 20 cm, 30-50% EtOAc in hexanes gradient
elution) afforded 3a. In some cases, repeated chromatography was required to
separate 3a from its undesired amino alcohol regioisomer and excess benzyl
carbamate.
J. Org. Chem. Vol. 73, No. 22, 2008 8977

Banerjee et al.
filtered through Celite (washed with MeOH) to afford the crude
amino alcohol, which was used directly in the next step.
aq sodium potassium tartrate solution (5 mL), then extracted with
CH2Cl2 (3 × 10 mL). The combined organic layers were dried
(Na2SO4) and concentrated in vacuo. Fifty percent of the crude
amine mixture was subjected to flash chromatography (SiO2, 1-4%
MeOH in CH2Cl2 gradient elution) to afford samples of the pure
syn and anti diastereomers suitable for chiral HPLC analysis after
further derivatization (either N-Cbz or N-Ac protection; see below).
The remaining 50% of the crude amine mixture was treated with
benzyl chloroformate (15 µL, 18.1 mg, 0.098 mmol), Na2CO3 (12.0
mg, 0.11 mmol), and THF (1 mL). The resulting mixture was stirred
at rt for 18 h, concentrated in vacuo, and purified by flash
chromatography (90:10 SiO2-KF,17 0-30% EtOAc in hexanes
gradient elution), affording carbamates 15 or 17 as white solids
that were mixtures of diastereomers. Spectral data for 15a-c and
17a-c were consistent with previously reported data.2
N-Benzyl 2-Benzyloxycarbonylamino-3-(4-methoxyphenyl)-4-
methylpentanamide (15a). syn-15a was obtained in 97% ee, as
analyzed by HPLC (Chiralcel OD-H, 98:2 hexane:i-PrOH, 0.7 mL/
min; tR ) 20.6 min (major), 28.6 min). anti-15a was obtained in
92% ee, as analyzed by HPLC (Chiralcel OD-H, 98:2 hexane:i-
PrOH, 1 mL/min; tR ) 25.0 min (major), 31.6 min).
N-Benzyl 2-Benzyloxycarbonylamino-4-methyl-3-phenylpen-
tanamide (15b). syn-15b was obtained in 74% ee, as analyzed by
HPLC of the derivative in which the N-Cbz group is replaced by
an acetate2 (Chiralcel OD-H, 98:2 hexane:i-PrOH, 1 mL/min; tR
) 6.2 min (major), 8.8 min). anti-15b was obtained in 87% ee, as
analyzed by HPLC under identical conditions (tR ) 18.1 min
(major), 22.9 min).
N-Benzyl 2-Benzyloxycarbonylamino-3-(4-fluorophenyl)-4-me-
thylpentamide (15c). syn-15c was obtained in 70% ee, as analyzed
by HPLC of the derivative in which the N-Cbz group is replaced
by an acetate2 (Chiralcel OD-H, 98:2 hexane:i-PrOH, 1 mL/min;
tR ) 5.9 min (major), 9.2 min). anti-15c was obtained in 63% ee,
as analyzed by HPLC under identical conditions (tR ) 16.2 min
(major), 21.2 min).
A solution of 41c (73.8 mg, 0.25 mmol) in anhydrous CHCl3
(2.0 mL) was stirred at 0 °C under Ar for 5 min, then treated
dropwise with a solution of the crude amino alcohol and Et3N (80
µL, 58.2 mg, 0.58 mmol) in anhydrous CHCl3 (1.0 mL). The
resulting mixture was stirred at 35 °C under Ar for 24 h, then treated
with solid NH4Cl (50 mg), stirred at rt for 30 min, and filtered.
The solid was stirred in THF (10 mL) for 30 min, and the mixture
was filtered. The combined organic solutions were concentrated in
vacuo. Flash chromatography (SiO2, 1.5 × 20 cm, 50-100% EtOAc
in hexanes gradient elution) afforded 5a (133.9 mg, 0.23 mmol,
91%) as a yellow oil: [R]25D +131 (c 0.10, EtOH); 1H NMR (CDCl3,
500 MHz) δ 8.10 (d, J ) 7.5 Hz, 2H), 8.03 (d, J ) 7.0 Hz, 4H),
7.86 (s, 2H), 7.80-7.72 (m, 8H), 7.50-7.47 (m, 4H), 7.42-7.39
(m, 2H), 5.45 (dd, J ) 10.0, 6.5 Hz, 2H), 4.07 (dd, J ) 12, 3.5 Hz,
2H), 3.99 (dd, J ) 11.5, 6.5 Hz, 2H), 3.51 (br s, 2H); 13C NMR
(CDCl3, 125 MHz) δ 164.5 (2C), 153.4 (2C), 136.5 (2C), 133.3
(2C), 132.9 (2C), 128.6 (2C), 127.9 (2C), 127.6 (2C), 127.5 (2C),
126.2 (2C), 125.9 (2C), 125.6 (2C), 124.8 (2C), 124.3 (2C), 124.2
(2C), 123.6 (2C), 118.9 (2C), 66.0 (2C), 56.6 (2C); IR (film) νmax
3325, 2920, 2850, 1731, 1695, 1682, 1658, 1641, 1592, 1547, 1539,
1531, 1462, 1060 cm-1; HRMS (ESI) m/z 595.2226 (MH+,
C38H30N2O5H requires 595.2228).
4,6-Bis((R)-4-(naphthalen-2-yl)-4,5-dihydrooxazol-2-yl)diben-
zo[b,d]furan (6a). A solution of 5a (59.2 mg, 0.10 mmol) in
anhydrous CH2Cl2 (1.5 mL) was treated with Et3N (40 µL, 29.1
mg, 0.29 mmol) and 4-pyrrolidinopyridine (4.8 mg, 0.032 mmol).
The mixture was stirred at 0 °C for 10 min, then treated dropwise
with a solution of TsCl (45.9 mg, 0.23 mmol) in anhydrous CH2Cl2
(1.5 mL). The resulting mixture was vigorously stirred at rt for
19 h, then treated with saturated aq NH4Cl (7 mL) and extracted
with CH2Cl2 (3 × 10 mL). The combined organic layers were dried
(Na2SO4) and concentrated in vacuo. Flash chromatography (SiO2,
1.5 × 19 cm, 20-100% EtOAc in hexanes gradient elution)
Methyl 2-Benzyloxycarbonylamino-3-(4-methoxyphenyl)-4-me-
thylpentanoate (17a). syn-17a was obtained in 10% ee, as analyzed
by HPLC (Chiralcel OD-H, 98:2 hexane:i-PrOH, 1 mL/min; tR )
17.4 min (major), 23.9 min). anti-17a was obtained in 7% ee, as
analyzed by HPLC under identical conditions (tR ) 11.8 min
(major), 23.0 min).
Methyl 2-Benzyloxycarbonylamino-4-methyl-3-phenylpentanoate
(17b). syn-17b was obtained in 7% ee, as analyzed by HPLC
(Chiralcel OD-H, 98:2 hexane:i-PrOH, 1 mL/min; tR ) 11.7 min
(major), 18.3 min). anti-17b was obtained in 5% ee, as analyzed
by HPLC under identical conditions (tR ) 7.3 min (major), 21.8
min).
afforded 6a (49.1 mg, 0.088 mmol, 88%) as an off-white solid:
1
[R]25 -28 (c 0.13, EtOH); H NMR (CDCl3, 500 MHz) δ 8.24
D
(d, J ) 8.0 Hz, 2H), 8.17 (d, J ) 7.5 Hz, 2H), 7.87 (s, 2H), 7.80
(d, J ) 8.0 Hz, 4H), 7.77 (d, J ) 8.5 Hz, 2H), 7.53-7.48 (m, 4H),
7.44-7.36 (m, 4H), 5.68 (t, J ) 9.0 Hz, 2H), 4.96 (t, J ) 9.0 Hz,
2H), 4.40 (t, J ) 8.0 Hz, 2H); 13C NMR (CDCl3, 125 MHz) δ
162.5 (2C), 154.4 (2C), 145.3 (2C), 139.8 (2C), 133.4 (2C), 132.8
(2C), 128.8 (2C), 128.6 (2C), 127.9 (2C), 127.7 (2C), 126.2 (2C),
125.8 (2C), 125.5 (2C), 124.9 (2C), 124.8 (2C), 123.9 (2C), 123.1
(2C), 74.8 (2C), 70.1 (2C); IR (film) νmax 2928, 1650, 1494, 1427,
1185, 1124, 984 cm-1; HRMS (ESI) m/z 559.2026 (MH+,
C38H26N2O3H requires 559.2016).
Methyl 2-Benzyloxycarbonylamino-3-(4-fluorophenyl)-4-meth-
ylpentanoate (17c). syn-17c was obtained in 3% ee, as analyzed
by HPLC (Chiralcel OD-H, 98:2 hexane:i-PrOH, 1 mL/min; tR )
11.5 min (major), 16.7 min). anti-17c was obtained in 3% ee, as
analyzed by HPLC under identical conditions (tR ) 7.6 min (major),
14.9 min).
General Procedure for Enantioselective Radical Conjugate
Additions Promoted by Mg(NTf2)2 and 6a. A solution of Mg(NTf2)2
(64.0 mg, 0.11 mmol) and ligand 6a (50.0 mg, 0.11 mmol) in
anhydrous CH2Cl2 (1.5 mL) was stirred at rt for 16 h, then treated
with 14 or 16 (0.11 mmol). The walls of the reaction vessel were
washed with CH2Cl2 (0.5 mL), and the mixture was stirred at -78
°C for 30 min. Isopropyl iodide (60 µL, 100 mg, 0.60 mmol),
Bu3SnH (72 µL, 78 mg, 0.27 mmol), Et3B (3.45 M solution in
CH2Cl2, 157 µL, 0.55 mmol), and O2 (5 mL) were added
sequentially, and identical quantities of these reagents were added
twice more at 1.5 h intervals. The mixture was stirred at -78 °C
for an additional 1 h (4 h total since initiation of radical reaction),
treated with 2 N HCl (5 mL), and extracted with CH2Cl2 (2 × 5
mL). The combined organic layers were concentrated in vacuo.
The crude adducts were treated with concd HCl (0.30 mL, 3.6
mmol), H2O (1 mL), THF (1 mL), and indium powder (101.0 mg,
0.88 mmol). The resulting mixture was stirred at rt for 15 h, the
solids were removed, and the volatiles were removed in vacuo.
The residue was diluted with 1 N HCl (5 mL), and tin byproducts
were removed by extraction with hexanes (5 × 5 mL). The aqueous
layer was treated with Na2CO3 (added until pH ∼8) and saturated
Acknowledgment. We thank Brigham Young University
(Cancer Research Center Fellowship to SGC) and the National
Institutes of Health (GM70483) for financial support.
Supporting Information Available: General experimental
details, synthetic procedures, spectral data, and 1H and 13C NMR
spectra for all new compounds, as well as HPLC traces for
enantioselective radical conjugate additions. This material is
available free of charge via the Internet at http://pubs.acs.
org.
JO801721Z
(17) Harrowven, D. C.; Guy, I. L. Chem. Commun. 2004, 1968.
8978 J. Org. Chem. Vol. 73, No. 22, 2008