An NMR strategy for determination of configuration of remote stereogenic centers: 3-methylcarboxylic acids

Article abstract of DOI:10.1021/ja972664n

A method has been developed for determining the absolute configuration of carboxylic acids bearing a methyl substituent at C(3). A series of 1-arylethylamide derivatives of such acids was prepared in which both the amine- and acid-derived portions were of

Full text of DOI:10.1021/ja972664n

4638
J. Am. Chem. Soc. 1998, 120, 4638-4643
An NMR Strategy for Determination of Configuration of Remote
Stereogenic Centers: 3-Methylcarboxylic Acids
Thomas R. Hoye* and Dmitry O. Koltun
Contribution from the Department of Chemistry, UniVersity of Minnesota, Minneapolis, Minnesota 55455
ReceiVed August 4, 1997
Abstract: A method has been developed for determining the absolute configuration of carboxylic acids bearing
a methyl substituent at C(3). A series of 1-arylethylamide derivatives of such acids was prepared in which
both the amine- and acid-derived portions were of known configuration. Diagnostic chemical shift differences
(∆δ’s) were identified for various proton resonances in each pair of diastereomeric amides and a new method
was established based on the observed trends. A conformational model (supported by computational
calculations) consistent with the observed differential shielding effects is offered. This approach represents a
general strategy that can be adapted to other substructures.
NMR-based methods, utilizing various chiral auxiliaries, for
R-Arylethylamides5 (as well as -carbamates and -ureas)
constitute the basis for NMR methods applicable to configu-
rational determination of R-branched carboxylic acids (as well
as secondary alcohols and amines) because of the well-
recognized conformational properties of the XC(dO)NHCH-
(Me)(Ar) substructural unit.6 The configurations of R-substi-
tuted acids have also been assigned from various phenyl-
glycinamide derivatives.7
No NMR method is available for determination of configu-
ration in remotely branched (i.e., â-, γ-, or δ-chiral) carboxylic
acids. We have evaluated the potential of different 1-arylethy-
lamines (2-5) as derivatizing agents for chiral carboxylic acids
and conclude that those derived from 1-phenylethylamine (2)
and 1-(1-naphthalenyl)ethylamine (3) are attractive for the
assignment of configuration of chiral acids 1e.
establishing the absolute configuration of commonly encoun-
tered structural units are well-known.1 Such methods are
valuable in the context of natural product structure determina-
tion, especially in situations where the quantity of material is
sufficiently limited to make degradation studies impractical. The
functional group handle that is derivatized by the auxiliary is
nearly always directly attached to the stereogenic center (i.e.,
R-branched) under examination (cf., 1a). In a few cases,
strategies have emerged for deducing the configuration of
substrates in which the stereocenter is remote to the functional
group handle. Strategies are available for â-branched primary
alcohols [1b; Mosher esters/Eu(fod)3],2 amine derivatives (1c;
Pirkle isocyanates),3 and aldehydes (1d; ephedrine-derived
oxazolidines).4
Acyclic carboxylic acids containing remote stereocenters
constitute another class of substrates for which a method of this
type would be useful. â-Substituted acids 1e represent a specific
structural unit that is of interest to us in the context of several
natural product structural questions. We describe here a solution
to this problem and offer insight to the underlying conforma-
tional issues that are responsible for the success of the method.
The approach also serves as a general strategy applicable to
additional structural types.
To determine whether there would be sufficient anisotropic
shielding to distinguish protons residing various distances from
the carboxyl group, we first prepared the homologous series of
isoamides derived from 2 through 5 and isobutyric (2-methyl-
propanoic), isovaleric (3-methylbutanoic), isocaproic (4-meth-
ylpentanoic), and 5-methylhexanoic acids. Amines 2 and 3 are
commercially available in optically pure form, and compounds
(1) (a) Yamaguchi, S. In Asymmetric Synthesis, Vol. 1, Analytical
Methods; Morrison, J. D., Ed.; Academic Press: New York, 1983; pp 125-
152. (b) Parker, D. Chem. ReV. 1991, 91, 1441. It is important to distinguish
between (the more numerous) methods that are suitable for determination
of enantiomeric excess and those useful for the more demanding task of
determination of configuration (e.g., Valentine, D., Jr.; Chan, K. K.; Scott,
C. G.; Johnson, K. K.; Toth, K.; Saucy, G. J. Org. Chem. 1976, 41, 62).
(2) Yasuhara, F.; Yamaguchi, S. Tetrahedron Lett. 1977, 4085 and
subsequent applications of that method.
(3) E.g.: (a) Pirkle, W. H.; Simmons, K. A. J. Org. Chem. 1983, 48,
2520. (b) Pirkle, W. H.; Robertson, M. R.; Hyun, M. H. J. Org. Chem.
1984, 49, 2433.
(4) Agami, C.; Meynier, F.; Berlan, J.; Besace, Y.; Brochard, L. J. Org.
Chem. 1986, 51, 73.
(5) E.g.: (a) Helmchen, G.; Vo¨lter, H.; Schu¨hle, W. Tetrahedron Lett.
1977, 1417 and references therein. (b) DeMunari, S.; Marazzi, G.; Forgione,
A.; Longo, A.; Lombardi, P. Tetrahedron Lett. 1980, 21, 2273.
(6) Pirkle, W. H.; Finn, J. In Asymmetric Synthesis, Vol. 1, Analytical
Methods; Morrison, J. D., Ed.; Academic Press: New York, 1983; pp 87-
124 and references therein.
(7) Nagai, Y.; Kusumi, T. Tetrahedron Lett. 1995, 36, 1853.
S0002-7863(97)02664-4 CCC: $15.00 © 1998 American Chemical Society
Published on Web 04/30/1998

Determining Configuration of Remote Stereogenic Centers
J. Am. Chem. Soc., Vol. 120, No. 19, 1998 4639
Table 1. ∆δ Values of the â-Methyl and Other Diagnostic
Resonances in Diastereomeric Syn and Anti 1-Arylethylamides from
3-Methylcarboxylic Acidsa
Figure 1. The magnitude of the anisotropic shielding effect (∆δ) for
diastereotopic Me groups in a series of R-, â-, γ-, and δ-branched
carboxylic acid, N-1-arylethylamides.
4 and 5 were prepared to perform the initial studies. The
1
differences in H NMR chemical shifts of the diastereotopic
methyl groups are shown in Figure 1. As expected, the
magnitude of the anisotropic effect for the anthracene-containing
derivatives of 5 is generally larger than that for the other
analogues. The second-strongest effect was observed for
1-naphthalenyl derivatives of 3, and smaller ∆δ values were
observed for 4 and 2.8,9 Nevertheless, all of the amide
derivatives provided sufficiently large differential anisotropic
shielding to distinguish the diastereotopic methyl groups in the
â-branched amides (cf. Figure 1, n ) 1). This suggests that
any of these amide derivatives could be used as chiral deriva-
tizing agents for â-branched chiral acids. We used optically
pure amines 2 and 3 for our further investigation because of
the convenience of their ready availability.
a In some cases the enantiomers of the compounds shown here were
prepared. These are designated as ent-## in the Experimental Section.
This, of course, does not affect the ∆δ values reported here for the
diastereomeric pairs. b Data from ref 11. The compounds were actually
the enantiomers of those shown here. c See ref 18.
It was first necessary to establish whether there would be a
reliable trend in both the magnitude and, more importantly, sign
of the chemical shift differences10 across a series of amides
derived from 3-substituted acids of known configuration. The
data from amides 6-11 (from the phenyl-containing amine 2)
and 12-16 (from the 1-naphthalenyl-containing amine 3) are
summarized in Table 1. In entry 1 are the ∆δ values for
diagnostic resonances in the diastereomeric syn- and anti-(R)-
N-1-phenylethylamides of S- and R-3-methylpentanoic acids (6s
and 6a, respectively). We use syn/anti to define the relative
orientation of the benzylic methyl group and the methyl
substituent at C(3) when the main chains of the amides are
oriented as shown in the structures at the top of Table 1. We
define ∆δ for any given resonance in a pair of diastereomeric
amides as the value of the chemical shift in the syn diastereomer
minus the value of the chemical shift of the anti diastereomer
(i.e., ∆δ ) δsyn - δanti). The ∆δ for the â-methyl group in the
diastereomers 6a/6s was positive (+0.032 ppm). The ∆δ of
the methyl protons of the ethyl group [i.e., C(5)] was both
opposite in sign (-0.012 ppm) and smaller in magnitude.
To account for the observed chemical shift differences, we
created a conformational model based on the following con-
siderations. The 1-arylethylamide derivatives used in our
method have numerous degrees of rotational freedom. Detailed
analysis of the conformational population is complex at best.
However, we have devised a working model, perhaps only a
mnemonic, to rationalize the observed shifts. The three
staggered rotamers i-iii arise from rotation about C(2)-C(3)
in the syn and anti 3-methylvaleramides 6s and 6a. With respect
to rotation about the N-Cbenzylic bond, we only consider those
(presumably dominant) conformations having the benzylic C-H
bond eclipsed with the amide carbonyl bond for the s-trans
amide rotamer (i.e., ω ) 180° and φC-H ) 0°). We assume
(8) Jacobus, J.; Raban, M.; Mislow, K. J. Org. Chem. 1968, 33, 1142.
(9) It is interesting that there appears to be a stepwise rather than linear
correlation between the rate of decease in the magnitude of the ∆δ values
and the number of methylenes in the isoamides. This suggests that an
“even-odd” effect may be operative.
(10) Comparison of magnitude and sign of ∆δ values between analogous
pairs of resonances in two diasteromeric derivatives is a commonly used
NMR strategy for deducing configuration and is most frequently encountered
in the Mosher method: (a) Dale, J. A.; Dull, D. L.; Mosher, H. S. J. Org.
Chem. 1969, 34, 2543. (b) Dale, J. A.; Mosher, H. S. J. Am. Chem. Soc.
1973, 95, 512. (c) Sullivan, G. R.; Dale, J. A.; Mosher, H. S. J. Org. Chem.
1973, 38, 2143. (d) Ohtani, I.; Kusumi, T.; Kashman, Y.; Kakisawa, H. J.
Am. Chem. Soc. 1991, 113, 4092.
the conformational populations among rotamers i-iii to be
similar for both diastereomers. Rotamers i and ii have the C(2)-
to-carbonyl carbon bond anti to either the C(3) ethyl or C(3)
methyl

4640 J. Am. Chem. Soc., Vol. 120, No. 19, 1998
Hoye and Koltun
group. Rotamers i and ii should be similarly populated. In
rotamer iii the C(2)-to-carbonyl carbon bond is anti to H(3) and
bisects the larger methyl and ethyl groups. Thus, rotamer iii
should be the least populated among i-iii. It follows that the
methyl attached to C(3) in the syn diastereomer 6s will be less
highly shielded relative to the anti diastereomer 6a. Since the
contribution of conformations iii is small, the relative shielding
of the C(3) methyl group is most dependent upon conformations
i, that of H(5) in the ethyl group upon conformation ii. Let us
examine how the other entries in Table 1 are also consistent
with this model.
Entry 2 shows data for the diastereomeric 1-phenylethyla-
mides derived form 3-cyclohexylbutanoic acid (7s and 7a),
which were recently described by Sibi and Porter.11 The C(3)
methyl ∆δ value is very similar in magnitude and identical in
sign with that of 6. Likewise, the ∆δ for the C(3)-methyl group
in the amides derived from (R)-citronellic acid 8s/8a was +0.033
(entry 3). The diastereomeric 3-methylglutarate derivatives 9s/
9a12 gave two readily distinguishable resonances. The ∆δ was
positive for the C(3)-methyl group (+0.013, entry 4) and
negative for the methoxy group (-0.015, entry 4). The
1-phenylethylamides of 3-methyl-4-pentenoic acids 10s/10a and
of 3,5-dimethyl-4-hexenoic acids 11s/11a show positive ∆δ
values for the C(3) methyl group and negative values for all of
the resonances associated with the olefin-containing substituents
(entries 5 and 6).13 The observed ∆δC(3)-Me values for all of
the diastereomeric pairs 6-11 are self-consistent and rational-
izable by the proposed conformational model. Namely, the
â-methyl groups appear farther downfield in the syn diastere-
omers than in the anti diastereomers. Where detectable, all
resonances associated with the other substituent at C(3) are
complementarily farther upfield in the syn isomers.
We have also prepared the analogous set of 1-(1-naphthale-
nyl)ethylamide derivatives 12-16. As expected, the magnitude
of the ∆δ value was larger for these naphthalenes by a factor
of ∼1.5-2 when compared with the phenyl-containing ana-
logues (cf. entries 7 vs 1, 8 vs 3, 9 vs 4, 10 vs 5, and 11 vs 6
in Table 1). The assignments of relative configuration for 12
and 13 rest on the known absolute configurations of the
precursor acid and amine (3). One of the diastereomers of 14
is known.12b,c The amides 15 were prepared from a racemic
sample of the acid and separated (MPLC, silica gel). We then
initially predicted the relative configuration of each isomer by
analogy of the sign of the ∆δ’s for both the C(3)-Me and the
vinyl resonances (entry 10) with the previous examples.
Subsequent hydrogenation of the sample assigned structure 15s
produced 12s, thereby proving its relative configuration and
further validating the method. Finally, the configuration of the
Figure 2. Arrays of the 10 lowest energy conformers representing the
majority of the population of 11s and 11a diastereomers of 3,5-dimethyl-
4-pentenoic acid (hydrogens removed for clarity) and their simplified
ChemDraw models. The circle represents the position of the â-methyl
substituent and the box that of the dimethylallyl moiety in 11.
1
16s/16a pair was assigned on the basis of analogous H NMR
spectral data and chromatographic behavior to those of 11s/
11a. Once again the â-methyl groups were always more highly
deshielded in the syn series than in the anti. Protons in the
other â-substituent were always relatively shielded in the syn
isomers, with one exception: we observed that the vinylic H(4)
in the 16s/16a pair actually has a positive, rather than the
expected negative, ∆δ value (entry 11). This suggests a caveat
for the application of this method; whenever possible, multiple
proton resonances corresponding to multiple sites in the
molecule should be analyzed.
To provide an alternative perspective of the conformational
issues in these amide derivatives, we have carried out a series
of Monte Carlo searches of conformational space using the
AMBER force field as implemented in the MacroModel14
software package. Due to the flexible nature of these acyclic
amides, dozens of conformations were found within 3 kcal/mol
from the global minimum energy conformation for each of the
3-substituted amides that was searched. No one of the calculated
low-energy conformations corresponds precisely to the rotamers
i-iii shown above, although distinct similarities can be found.
Instead of relying heavily on any single structure obtained from
the calculations, we consider families of low-energy conforma-
tions that represent the majority of the population. Such analysis
suggests those portions of the molecule that are likely to spend
more time in the shielding region of the aromatic ring.
(11) Sibi, M. P.; Ji J. G.; Wu, J. H.; Gurtler S.; Porter N. A. J. Am.
Chem. Soc. 1996, 118, 9200.
(12) (a) Assignment of the relative configuration in the major and minor
products arising from opening of 3-methylglutaric anhydride with 1-phe-
nylethylamine (2) (i.e., 9s and 9a, respectively) is based upon analogy with
formation of the diastereomeric pair 14s and 14a by opening with 1-(1-
naphthyl)ethylamine (3). Compound 14a has been previously reported12b,c
and its 1H NMR data matched those of our minor diastereomer {δ ) 3.638
(CO2Me) and δ ) 0.990 [C(3)Me]}. It is interesting that the sense of
diastereoselectivity for this amide formation is opposite that observed for
the reaction of 3-substituted glutaric anhydrides with 1-(1-naphthyl)-
ethanol12d but the same as that observed with mandelate esters.12b (b)
Konioke, T.; Araki, Y. J. Org. Chem. 1994, 59, 7849. (c) Harusawa, S.;
Takemura, S.; Yoneda, R.; Kurihara, T. Tetrahedron 1993, 49, 10577. (d)
Theisen, P. D.; Heathcock, C. H. J. Org. Chem. 1993, 58, 142.
As an example, consider the arrays of superpositions of the
10 lowest energy conformations of 11s and of 11a that are
shown in Figure 2. The C-(CdO)-NH-C sequence of atoms
was superimposed in all 10 structures to generate these views.
It is clear that in the family of low-energy conformers for 11s,
the phenyl ring is located quite near the 1-isobutenyl group in
9 out of the 10 lowest energy conformations. The opposite is
observed for the anti isomer 11a; the phenyl group is shielding
the â-methyl substituent in all 10 conformations. This same
trend is observed even when one examines families of, e.g., 20
(13) The stereochemical assignments for the 10s and 10a were proven
by hydrogenation of each separate diastereomer to authentic 6s and 6a,
respectively. An analogous proof was performed for each of 15s and 15a
to give 12s and 12a. The assignments for 11s/11a and 16s/16a rest on
(14) Mohamadi, F.; Richards, N. G. J.; Guida, W. C.; Liskamp, R.;
Lipton, M.; Caufield, C.; Chang, G.; Hendrickson, T.; Still, W. C. J. Comput.
Chem. 1990, 11, 440.
1
analogy of their H NMR spectral data Vis-a`-vis those of the 10s/10a and
15s/15a pairs.

Determining Configuration of Remote Stereogenic Centers
J. Am. Chem. Soc., Vol. 120, No. 19, 1998 4641
5H), 6.00 (br s, 1H), 5.12 (quintet, J ) 7.0 Hz, 1H), 2.17 (nfom, 1H),
1.95-1.83 (m, 2H), 1.45 (d, J ) 7.0 Hz, 3H), 1.34 (ddq, J ) 14, 7,
and 7 Hz, 1H), 1.17 (ddq, J ) 14, 7, and 7.0 Hz, 1H), 0.91 (d, J ) 6.0
Hz, 3H), and 0.86 (t, J ) 7.2 Hz, 3H). 13C NMR (DEPT) (75 MHz)
δ 171.9, 143.3, 128.6 (CH), 127.3 (CH), 126.2 (CH), 48.7 (CH), 44.1
(CH2), 32.4 (CH), 29.4, (CH2), 21.8 (CH3), 19.1 (CH3), and 11.3 (CH3).
IR (CHCl3) 3282 (broad), 3063, 3030, 2962, 2929, 1639, 1545, 1495,
and 1453 cm-1. LRMS (EI) m/z 219 (M+). Anal. Calcd: C, 76.66;
H, 9.65. Found: C, 76.82; H, 9.44.
or 30 of the calculated lowest energy conformations. For each
of 11s and 11a the energy of the 30th conformer was ∼1 kcal/
mol above the global minimum. These analyses are consistent
1
with the trends observed by H NMR spectroscopy for com-
pounds 11s/11a (Table 1, entry 6). We have also carried out a
similar force field conformational search with 6s/6a and
observed entirely analogous behavior. Thus, the computational
results reinforce the theory and support the assignment of
relative configuration in 11s and 11a. We recommend the use
of this type of routinely accessible computational strategy when
applying this methodology to new substrates.
[R-(R*,R*)]-3-Methyl-N-[1-phenylethyl]pentanamide (6a). The
same procedure as for 6s was used to convert a racemic mixture of
3-methylvaleric acid into a mixture of diastereomeric amides 6s and
6a (91%). Spectral data for the mixture (methyl resonances for 6a are
italicized): 1H NMR (300 MHz) δ 7.40-7.21 (m, 5H, Ph), 5.91 (br s,
1H), 5.13 (quintet, J ) 7.0 Hz, 1H), 2.17 (nfom, 1H), 1.95-1.83 (m,
2H), 1.48 (d, J ) 7.0 Hz, 3H), 1.41-1.14 (m, 2H), 0.91 (d, J ) 6.0
Hz, 1.5H), 0.88 (d, J ) 6.0 Hz, 1.5H), 0.87 (t, J ) 7.2 Hz, 1.5H), and
0.86 (t, J ) 7.2 Hz, 1.5H). 13C NMR (75 MHz) δ 171.9, 143.4/143.3,
128.6, 127.3, 126.2, 48.6, 44.2+/44.2-, 32.4+/32.4-, 29.4, 21.7, 19.2/
19.1, and 11.4/11.3. LRMS (EI) m/z 219 (M+).
In summary, we have developed a reliable method for
determining the absolute configuration of carboxylic acids
containing a stereogenic center at C(3). To apply this method
(1) identify the resonance of the â-Me group as well as
1
distinguishable H NMR resonances for protons unique to the
other substituent at C(3) (i.e., R in 17) in the diastereomeric
pair of 1-arylethylamides and (2) deduce the C(3)-configuration
by comparing the sign of ∆δ for one or more of these resonances
(Table 1 can be used as a convenient guide). For example, in
the anti isomer 17 of the R-amide of generic 3-methyl alkanoic
acids, the methyl resonance will be observed at higher field
than in the syn-diastereomer. Finally, complementarity exists;
that is, the ∆δ’s of resonances within R and those of the methyl
resonance will be of opposite sign. This approach to determin-
ing the configuration of a remote stereogenic center represents
a general strategy that can be adapted to other substructures.
[R-(R*,R*)]-3,7-Dimethyl-N-[1-phenylethyl]-6-octenamide (8a).
DCC coupling analogous to the one used for 6s was used to prepare
8a from commercially available R-citronellic acid and R-2. The crude
amide was purified by MPLC (4:1 hexanes/ethyl acetate) to give 8a as
white crystals (75%). Mp 62-63 °C. 1H NMR (500 MHz) δ 7.38-
7.20 (m, 5H), 5.95 (br s, 1H), 5.14 (quintet, J ) 7.0 Hz, 1H), 5.07 (t,
J ) 7.0 Hz, 1H), 2.19 (dd, J ) 5.0 and 13.0 Hz, 1H), 2.04-1.90 (m,
4H), 1.67 (s, 3H), 1.59 (s, 3H), 1.47 (d, J ) 7.0 Hz, 3H), 1.35 (m,
1H), 1.19 (m, 1H), and 0.90 (d, J ) 6.0 Hz, 3H). 13C NMR (DEPT,
HETCOR) (125 MHz) δ 171.6 (C, none), 143.4 (C, none), 131.4 (C,
none), 128.6 (CH, 7.3), 127.28 (CH, 7.2), 126.22 (CH, 7.3), 124.4 (CH,
5.07), 48.5 (CH, 5.14), 44.5 (CH2, 2.19, 1.92), 36.9 (CH2, 1.47, 1.35),
30.5 (CH, 1.88), 25.7 (CH3, 1.67), 25.4 (CH2, 2.00), 21.7 (CH3, 1.47),
19.5 (CH3, 0.90), and 17.68 (CH3, 1.59). Anal. Calcd: C, 79.07; H,
9.95. Found: C, 79.15; H, 9.81.
[R-(R*,S*)]-3,7-Dimethyl-N-[1-phenylethyl]-6-octenamide (ent-
8s). DCC coupling analogous to the one used for 6s was used to prepare
ent-8s from commercially available R-citronellic acid and S-2. The
crude amide was purified by MPLC (4:1 hexanes/ethyl acetate) to give
ent-8s as a colorless oil (65%). 1H NMR (500 MHz) δ 7.40-7.20 (m,
5H), 5.85 (br s, 1H), 5.14 (dq, J ) 7.0 and 7.0 Hz, 1H), 5.07 (t septets,
J ) 7.0 and 1.5 Hz, 1H), 2.18 (dd, J ) 4.5 and 13.0 Hz, 1H), 2.10-
1.90 (m, 4H), 1.66 (s, 3H), 1.58 (s, 3H), 1.48 (d, J ) 7.0 Hz, 3H),
1.31-1.39 (m, 1H), 1.17-1.21 (m, 1H), and 0.93 (d, J ) 6.0 Hz, 3H).
13C NMR (125 MHz) δ 171.61, 143.30, 131.45, 128.64, 127.32, 126.23,
124.37, 48.57, 44.57, 36.92, 30.55, 25.72, 25.44, 21.71, 19.51, and
17.66.
Experimental Section
General Procedures and Methods. (S)-6-Methoxy-R-methyl-2-
naphthalenemethanamine (4) was prepared from commercial (+)-(S)-
naproxen via the Curtius rearrangement route by a procedure analogous
to one described in the literature for the Shioiri reaction using
(PhO)2PON3.15 The amides indicated in Figure 1 that were derived
from isoacids (isobutyric through 5-methylhexanoic) and amines 2
through 5 were prepared via standard coupling procedures (DCC or
the acid chloride) on a small scale for the practical purpose of obtaining
1H NMR spectra. 9-Anthracenyl-derived amine 5 was obtained
following the literature procedure.16 3-Methylpent-4-enoic acid (the
precursor for 10s/10a and 15s/15a) and 3,5-dimethylhex-4-enoic acid
(the precursor for 11s/11a and 16s/16a) were noncommercial samples
available in our laboratory and prepared by Claisen rearrangement. All
1H and 13C NMR spectra were measured as CDCl3 solutions.
[S-(R*,S*)]-3-Methyl-N-[1-phenylethyl]pentanamide (6s). In a 5
mL culture tube dicyclohexylcarbodiimide (DCC) (83 mg, 0.4 mmol)
and (dimethylamino)pyridine (DMAP) (5 mg) were dissolved in 0.5
mL of dry CH2Cl2. (S)-3-Methylvaleric acid17 (50 µL, 0.4 mmol) was
added neat and a thick white precipitate formed instantaneously. After
the mixture was stirred for 5 min (R)-R-methylbenzylamine (50 µL,
0.39 mmol) was added neat. The reaction mixture was stirred for 2 h
and filtered through a layer of silica gel. The liquid was concentrated
and purified by MPLC (3:1 hexanes/ethyl acetate) to give a colorless
viscous oil (48.6 mg, 55%). 1H NMR (300 MHz) δ 7.40-7.22 (m,
[R-(R*,R*)]- and [S-(R*,S*)]-3-Methyl-5-[(1-phenylethyl)amino]-
5-oxopentanoic Acid, Methyl Ester (9s and 9a). 3-Methylglutaric
anhydride (101.2 mg, 0.79 mmol), 1-(R)-phenylethylamine (101 µL,
0.79 mmol), and DMAP (100 mg, 0.82 mmol) were dissolved in 2 mL
of dry CH2Cl2 and stirred for 1 h. DCC (163 mg, 0.79 mmol) was
added as one portion. After several minutes 1 mL of dry MeOH was
added. The reaction mixture was stirred for 24 h, then filtered through
silica. The solution was washed with 5% HCl (2 × 2 mL) and
concentrated NaHCO3 (2 mL). Drying over MgSO4 and concentration
in vacuo yielded compounds 9s and 9a as a mixture with a 1.9:1 ratio
(colorless oil) (191.6 mg, 92%). The crude material was of sufficient
purity for spectroscopic analysis, with only a trace of 3-methylglutaric
acid dimethyl ester as an impurity. The NMR resonances of 9s vs 9a
were easy to distinguish because of the difference in relative intensity.
9s: 1H NMR (500 MHz) δ 7.38-7.21 (m, 5H), 5.90 (brs, 1H), 5.16
(18) The assignment of resonances to the methyl groups (6Z and 6E)
for the 11s/11a and 16s/16a pairs is based both on their relative chemical
shifts [cf. other 1,1-dimethyl-substituted alkenes in which the cis-methyl
group appears upfield from the trans-methyl group (e.g., the six cis-disposed
methyl groups in squalene have a chemical shift of 1.60 ppm while the
two trans-disposed methyls are at 1.68 ppm)]. Even if the differential
anisotropic shielding effect of the aromatic ring in either of these pairs
were sufficiently large to reverse that assignment for one or both of the
diastereomers, the signs of the resulting ∆δ values would remain negative.
(15) (a) Shishido, K.; Shitara, E.; Komatsu, H.; Hiroya, K.; Fukumoto,
K.; Kametani, T. J. Org. Chem. 1986, 51, 3007. (b) Wolber, E. K. A.;
Ru¨chardt, C. Chem. Ber. 1991, 124, 1667.
(16) Ciganek, E., U.S. Patent 4 076 830, 1978; Chem. Abstr. 89, 24136.
(17) An authentic sample of (S)-3-methylvaleric acid was prepared by
the deamination of L-isoleucine: Doldouras, G. A.; Kollonitsch, J. J. Am.
Chem. Soc. 1978, 100, 341.

4642 J. Am. Chem. Soc., Vol. 120, No. 19, 1998
Hoye and Koltun
(dq, J ) 7.0 and 7.0 Hz, 1H), 3.66 (s, 3H), 2.45-2.20 (m, 4H), 2.10
(dd, 1H, J ) 14.0 and 7.0 Hz), 1.44 (d, 3H, J ) 7.0 Hz), and 1.02 (d,
MHz) δ 1H NMR (CDCl3, 500 MHz) δ 8.10 (d, 1H, J ) 7.5 Hz), 7.86
(dd, 1H, J ) 7.0 and 2.2 Hz), 7.79 (d, 1H, J ) 8.0 Hz), 7.42-7.55 (m,
4H), 5.95 (dq, 1H, J ) 6.9 and 6.9 Hz), 5.69 (br d, 1H, J ) 7.0 Hz),
2.15 (m, 1H), 1.94-1.82 (m, 2H), 1.66 (d, 3H, J ) 6.6 Hz), 1.34 (m,
1H), 1.16 (m, 1H), 0.92 (d, 1.5H, J ) 6.0 Hz), 0.87 (t, 1.5H, J ) 7.2
Hz), 0.87 (d, 1.5H, J ) 6.0 Hz), and 0.84 (t, 1.5H, J ) 7.2 Hz). 13C
NMR (DEPT) (75 MHz) δ 171.44 (C), 138.22 (C), 133.90 (C), 131.17
(C), 128.70 (CH), 128.37 (CH), 126.50 (CH), 125.89 (CH), 125.12
(CH), 123.61 (CH), 122.56 (CH), 44.37 (CH3), 44.24 (CH2), 32.40
(CH), 29.43 (CH2), 20.55 (CH3), 19.19 (CH3), and 11.33 (CH3). LRMS
(EI) m/z 269 (M•+), 254, 213, 198, 170, 156 (100%), 142, 129, 115,
99, 85, 81, 71, and 57. FT-IR 3439, 1653 cm-1. Anal. Calcd: C,
80.25; H, 8.61. Found: C, 80.12; H, 8.63.
1
7.0 Hz). The H NMR spectrum of 9a was virtually the same as that
for 9s with the following differences: δ 3.67 (s, 3H) and 1.01 (d, 7.0
Hz). 13C NMR (DEPT) (9a/9s mixture) (75 MHz) δ 173.14 (C), 170.67
(C), 143.34 (C), 128.62 (CH), 127.28 (CH), 126.14 (CH), 51.49/51.29
(CH3), 48.64 (CH), 43.03 (CH2), 40.35/40.30 (CH2), 28.25 (CH), 21.88/
21.79 (CH3), and 19.29 (CH3). LRMS (EI) (9a/9s mixture) m/z 263
(M•+). FT-IR (9a/9s mixture) 3436, 1728, 1659 cm-1. Anal. Calcd:
C, 68.42; H, 8.04. Found: C, 68.23; H, 8.05.
[R-(R*,R*)]-3-Methyl-N-[1-phenylethyl]-4-pentenamide (10s) and
[S-(R*,S*)]-3-Methyl-N-[1-phenylethyl]-4-pentenamide (10a). (()-
3-Methyl-4-pentenoic acid was coupled by using DCC (cf. preparation
of 6s) with R-2 to give a mixture of diastereomers 10s/10a that was
separated by MPLC (4:1 hexanes/ethyl acetate) to give 10s and 10a as
individual compounds (>95% de by NMR). 10s: 1H NMR (500 MHz)
δ 7.40-7.20 (m, 5H), 5.76 (ddd, J ) 6.9, 10.2, and 17.1 Hz, 1H), 5.70
(br s, 1H), 5.15 (dq, 1H, J ) 7.5 and 7.5 Hz), 5.01 (dd, J ) 1.5 and
17.4 Hz, 1H), 4.94 (dd, J ) 1.2 and 10.2 Hz, 1H), 2.69 (septet, J )
6.9 Hz, 1H), 2.25-2.08 (m, 2H), 1.48 (d, J ) 6.6 Hz, 3H), and 1.05
(d, J ) 6.6 Hz, 3H). 10a: 1H NMR (500 MHz) δ 7.37-7.18 (m, 5H),
5.77 (ddd, J ) 6.9, 10.2, and 17.1 Hz, 1H), 5.67 (br s, 1H), 5.15 (dq,
1H, J ) 7.5 and 7.5 Hz), 5.03 (dd, J ) 1.5 and 17.4 Hz, 1H), 4.97 (dd,
J ) 1.2 and 10.2 Hz, 1H), 2.70 (septet, J ) 6.9 Hz, 1H), 2.24-2.08
(m, 2H), 1.47 (d, J ) 6.6 Hz, 3H), and 1.03 (d, J ) 6.6 Hz, 3H).
10s/10a mixture: 13C NMR (DEPT) (75 MHz) δ 170.81 (C), 143.24
(C), 142.80 (CH), 128.60/128.58 (CH), 127.28 (CH), 126.22/126.20
(CH), 113.52 (CH2), 48.62/48.59 (CH), 43.83 (CH2), 34.91/34.85 (CH),
21.73/21.69 (CH3), 19.69/19.64 (CH3). LRMS (EI): m/z 217 (M•+),
202, 188, 174, 162, 120, 106, 105 (100%), 98, 91, 77, 69 and 55. FT-
IR 3437, 1658 cm-1. Anal. Calcd: C, 77.38; H, 8.81. Found: C,
77.39; H, 8.83.
[S-(R*,R*)]-3,5-Dimethyl-N-[1-phenylethyl]-4-hexenamide (ent-
11s) and [R-(R*,S*)]-3,5-Dimethyl-N-[1-phenylethyl]-4-pentenamide
(ent-11a). (()-3,5-Dimethyl-4-hexenoic acid was coupled by using
DCC (cf. preparation of 6s) with S-2 to give a mixture of diastereomers
ent-11s/ent-11a that was separated by MPLC (4:1 hexanes/ethyl acetate)
to give ent-11s and ent-11a as individual compounds (>95% de by
NMR). The relative (syn or anti) configuration was assigned on the
basis of a trend in chemical shift differences (δ∆’s) observed for
diastereomeric amides 6-10 and 12-15 and the results of computa-
tional studies (see text). ent-11s: 1H NMR (500 MHz) δ 7.21-7.39
(m, 5H), 5.70 (br d, 1H, J ) 7.4 Hz), 5.11 (dq, 1H, J ) 7.1 and 7.1
Hz), 4.91 (d of septets, 1H, J ) 9.7 and 1.3 Hz), 2.85 (dtq, 1H, J )
5.7, 9.2 and 6.8 Hz), 2.17 (dd, 1H, J ) 5.5 and 13.8 Hz), 2.06 (dd, 1H,
J ) 9.0 and 14.1 Hz), 1.62 (d, 3H, J ) 1.5 Hz), 1.52 (d, 3H, J ) 1.5
Hz), 1.47 (d, 3H, J ) 7.0 Hz), and 0.97 (d, 3H, J ) 6.5 Hz). LRMS
(EI) m/z 245 (MI•+), 230, 202, 174, 163, 148, 120, 105 (100%), 83,
59, 55. ent-11a: 1H NMR (500 MHz) δ 7.21-7.39 (m, 5H), 5.71 (br
d, 1H, J ) 7.7 Hz), 5.11 (dq, 1H, J ) 7.1 and 7.1 Hz), 4.93 (d of
septets, 1H, J ) 9.7 and 1.5 Hz), 2.86 (dtq, 1H, J ) 5.8, 9.0 and 6.4
Hz), 2.16 (dd, 1H, J ) 5.8 and 14.1 Hz), 2.07 (dd, 1H, J ) 8.8 and
14.0 Hz), 1.67 (d, 3H, J ) 1.0 Hz), 1.64 (d, 3H, J ) 1.0 Hz), 1.44 (d,
3H, J ) 7.0 Hz), and 0.96 (d, 3H, J ) 6.5 Hz). LRMS (EI) identical
with that of ent-11s.
[R-(R*,S*)]-3,7-Dimethyl-N-[1-(1-naphthalenyl)ethyl]-6-octena-
mide (ent-13s) and [R-(R*,R*)]-3,7-Dimethyl-N-[1-(1-naphthalenyl)-
ethyl]-6-octenamide (13a). (R)-Citronellic acid was coupled by using
DCC (cf. preparation of 6s) with R-3 and S-3 to give diastereomers
1
13a and ent-13s correspondingly. 13a: H NMR (500 MHz) δ 8.07
(d, 1H, J ) 7.0 Hz), 7.85 (dd, 1H, J ) 7.0 and 2.0 Hz), 7.79 (d, 1H,
J ) 7.0 Hz), 7.41-7.55 (m, 4H), 5.95 (dq, J ) 7.0 and 7.0 Hz, 1H),
5.85 (br d, 1H, J ) 7.0 Hz), 5.04 (t septets, J ) 7.0 and 1.5 Hz, 1H),
2.15-1.90 (m, 5H), 1.66 (d, J ) 6.3 Hz, 3H), 1.65 (s, 3H), 1.56 (s,
3H), 1.40-1.05 (m, 2H), and 0.87 (d, J ) 6.3 Hz, 3H). The 1H NMR
spectrum of ent-13s was virtually identical with that of 13a with the
exception of the aliphatic methyl resonance: δ 0.92 (d, 3H, J ) 6.3
Hz).
[R-(R*,R*)]- and [S-(R*,S*)]-3-Methyl-5-[[1-(1-naphthalenyl)-
ethyl]amino]-5-oxopentanoic Acid, Methyl Ester (14s and 14a). The
compounds 14s and 14a were obtained from 3-methylglutaric anhydride
and R-3 by a procedure similar to the one used for the preparation of
compounds 9s/9a as a mixture with a 1.65:1 ratio. The NMR
resonances between these two were easy to distinguish by using the
difference in relative intensity. 14s: 1H NMR (500 MHz) δ 8.09 (d,
1H, J ) 7.8 Hz), 7.86 (dd, 1H, J ) 7.1 and 2.0 Hz), 7.80 (d, 1H, J )
8.0 Hz), 7.54-7.41 (m, 4H), 5.97-5.80 (m, 2H), 3.61 (s, 3H), 2.47-
2.01 (m, 5H), 1.65 (d, 3H, J ) 6.5 Hz), and 1.01 (d, 7.0 Hz). The 1H
NMR spectrum of 14a was virtually the same as for 14s, with the
following differences: δ 3.64 (s, 3H) and 0.99 (d, 7.0 Hz). Professor
Harusawa confirmed for us (7-30-96) that his sample of 14a had
resonances at δ 3.64 and 0.99 ppm and his sample of 14s had resonances
at δ 3.61 and 1.01 for the methyl ester and C(3) methyl groups,
respectively. In the original report these values were provided
ambiguously.12c
[S-(R*,R*)]-3-Methyl-N-[1-(1-naphthalenyl)ethyl]-4-pentena-
mide (ent-15s) and [R-(R*,S*)]-3-Methyl-N-[1-(1-naphthalenyl)-
ethyl]-4-pentenamide (ent-15a). (()-3-Methyl-4-pentenoic acid was
coupled by using DCC (cf. preparation of 6s) with S-3 to give a mixture
of diastereomers ent-15s/ent-15a that was separated by MPLC (4:1
hexanes/ethyl acetate) to give ent-15s and ent-15a as individual
compounds (>95% de by NMR). 13C NMR (ent-15s/ent-15a mixture)
(75 MHz) δ 170.55, 142.72, 138.17, 133.90, 131.14, 128.71, 128.38,
126.52, 125.89, 125.12, 123.60, 122.57, 113.53, 44.51/44.47, 43.84/
43.80, 34.88/34.83, 20.58, 19.68/19.64. LRMS (EI) (ent-15s/ent-15a
mixture) m/z 267 (M•+). IR: 3438 and 1657 cm-1. This mixture was
fractionated (MPLC, silica gel, 4:1 hexanes/ethyl acetate) into enriched
samples of ent-15s (92% de by NMR) and ent-15a (80% de by NMR).
ent-15s: 1H NMR (500 MHz) δ 8.10 (d, 1H, J ) 8.5 Hz), 7.86 (dd,
1H, J ) 8.0 and 1.5 Hz), 7.80 (d, 1H, J ) 8.0 Hz), 7.44-7.55 (m,
4H), 5.95 (dq, 1H, J ) 7.0 and 7.0 Hz), 5.75 (ddd, 1H, J ) 7.0, 10.5,
and 17.5 Hz), 5.66 (br d, 1H, J ) 7.0 Hz), 5.02 (ddd, 1H, J ) 1.5, 1.5,
and 17.5 Hz), 4.95 (ddd, 1H, J ) 1.5, 1.5, and 10.5 Hz), 2.72 (∼septet,
1H, J ∼ 6.5 Hz), 2.18 (dd, 1H, J ) 6.5 and 14.0 Hz), 2.08 (dd, 1H, J
) 6.0 and 14.0 Hz), 1.66 (d, 3H, J ) 6.5 Hz), and 1.01 (d, 3H, J ) 6.5
Hz). ent-15a: 1H NMR (500 MHz) δ 8.09 (d, 1H, J ) 8.0 Hz), 7.87
(dd, 1H, J ) 7.5 and 2.0 Hz), 7.80 (d, 1H, J ) 8.0 Hz), 7.43-7.55 (m,
4H), 5.95 (dq, 1H, J ) 7.0 and 7.0 Hz), 5.72 (ddd, 1H, J ) 7.0, 10.5,
and 17.5 Hz), 5.66 (br d, 1H, J ) 7.0 Hz), 4.97 (ddd, 1H, J ) 1.5, 1.5,
and 17.0 Hz), 4.90 (ddd, 1H, J ) 1.5, 1.5, and 10.5 Hz), 2.68 (∼septet,
1H, J ∼ 7.0 Hz), 2.19 (dd, 1H, J ) 7.5 and 14.0 Hz), 2.10 (dd, 1H, J
) 7.0 and 14.0 Hz), 1.67 (d, 3H, J ) 6.5 Hz), and 1.05 (d, 3H, J ) 7.0
Hz). Anal. Calcd: C, 80.86; H, 7.92. Found: C, 80.44; H, 7.83.
[S-(R*,R*)]-3-Methyl-N-[1-(1-naphthalenyl)ethyl]pentanamide (ent-
12a). By the method used to prepare 6s, (S)-3-methylvaleric acid and
S-3 were coupled with DCC to prepare ent-12a. Gradient elution
through a SiO2 column with hexanes and ethyl acetate gave a white
solid. 1H NMR (300 MHz), δ 8.11 (d, 1H, J ) 7.5 Hz), 7.87 (dd, 1H,
J ) 7.0 and 2.2 Hz), 7.80 (d, 1H, J ) 8.0 Hz), 7.4-7.55 (m, 4H), 5.96
(dq, 1H, J ) 6.9 and 6.9 Hz), 5.70 (br d, 1H, J ) 7.0 Hz), 2.16 (m,
1H), 1.9-1.8 (m, 2H), 1.67 (d, 3H, J ) 6.6 Hz), 1.35 (m, 1H), 1.17
(m, 1H), 0.87 (t, 3H, J ) 7.3 Hz), and 0.87 (d, 3H, J ) 6.3 Hz).
[R-(R*,S*)]-3-Methyl-N-[1-(1-naphthalenyl)ethyl]pentanamide (ent-
12s). The same procedure as for ent-12a was used to convert a racemic
mixture of 3-methylvaleric acid and S-3 into a mixture of diastereomeric
amides ent-12s and ent-12a (white solid, mp 120-126 °C) that gave
no indication of separation, even in the leading and trailing peak edges,
by MPLC on SiO2 (9:1 hexanes:EtOAc). Spectral data for the mixture
(the methyl resonances due to ent-12s are in italics): 1H NMR (300

Determining Configuration of Remote Stereogenic Centers
J. Am. Chem. Soc., Vol. 120, No. 19, 1998 4643
[S-(R*,R*)]-3,5-Dimethyl-N-[1-(1-naphthalenyl)ethyl]-4-hexena-
mide (ent-16s) and [R-(R*,S*)]-3,5-Dimethyl-N-[1-(1-naphthalenyl)-
ethyl]-4-pentenamide (ent-16a). (()-3,5-Dimethyl-4-hexenoic acid
was coupled by using DCC (cf. preparation of 6s) with S-3 to give a
mixture of diastereomers ent-16s/ent-16a that was separated by MPLC
(4:1 hexanes/ethyl acetate) to give ent-16s and ent-16a as individual
compounds (>95% de by NMR). The relative (syn or anti) configu-
ration was assigned on the basis of a trend in chemical shift differences
(δ∆’s) observed for diastereomeric amides 6-10 and 12-15 and the
results of computational studies (see text). ent-16s: 1H NMR (300
MHz) δ 8.09 (d, 1H, J ) 2.5 Hz), 7.87-7.78 (m, 2 Hz), 7.55-7.42
(m, 4H), 5.91 (dq, 1H, J ) 7.8 and 7.8 Hz), 5.67 (br d, 1H, J ) 7.5
Hz), 4.86 (d of septets, 1H, J ) 9.6 and 1.2 Hz), 2.85-2.75 (m, 1H),
2.17-2.04 (m, 2H), 1.64 (d, 3H, J ) 6.6 Hz), 1.53 (d, 3H, J ) 1.5
Hz), 1.40 (d, 3H, J ) 1.2 Hz), and 0.95 (d, 3H, J ) 6.6 Hz). LRMS
(EI) m/z 245 (MI•+), 230, 202, 174, 163, 148, 120, 105 (100%), 83,
59, 55. ent-16a: 1H NMR (300 MHz) δ 8.11 (d, 1H, J ) 2.5 Hz),
7.88-7.79 (m, 2 Hz), 7.56-7.42 (m, 4H), 5.91 (dq, 1H, J ) 7.8 and
7.8 Hz), 5.70 (br d, 1H, J ) 7.8 Hz), 4.84 (d of septets, 1H, J ) 9.6
and 1.5 Hz), 2.92-2.82 (m, 1H), 2.17 (dd, 1H, J ) 6.0 and 14.1 Hz),
2.06 (dd, 1H, J ) 8.4 and 14.1 Hz), 1.62 (d, 3H, J ) 6.6 Hz), 1.59 (d,
3H, J ) 1.2 Hz), 1.55 (d, 3H, J ) 1.5 Hz), and 0.93 (d, 3H, J ) 6.9
Hz). LRMS (EI) identical with that for ent-16s.
Acknowledgment. This work was supported by Grant No.
GM 39339 awarded by the DHHS. We thank Professors T.
1
Kurihara and S. Harusawa for providing H NMR spectra of
14a and 14s and Mr. Dean G. Brown for his valuable advice in
computational modeling.
JA972664N