contribution is the one generated on H(2) by the MPA unit
at C(3).6d
The above studies introduce four clearly distinct aniso-
tropic scenarios, one for each stereochemical situation. This
allows us to propose an unambiguous strategy to distinguish
them by means of two different NMR parameters related to
H(2) and H(3). This is summarized in the following points:
(A) The sign of the ∆δRS parameter for H(3) becomes
highly diagnostic7 and reduces the configurational possibili-
ties of a triol to just two out of four possible stereoisomers:
(1) If the ∆δRS of H(3) is positive, the stereochemistry of
the triol is either syn type A or anti type D. (2) If the ∆δRS
of H(3) is negative, the triol is either syn type B or anti type
C.
The validity of these predictions was experimentally
demonstrated from the 1H NMR spectra of the tris-(R)- and
the tris-(S)-MPA ester derivatives of triols 1-24 of known
absolute configuration (Figure 2). All the syn type A (1-5,
24) and the anti type D triols (21-23) present positive ∆δRS
for H(3), while those belonging to syn B (6-12) and anti C
(13-20, 24) present negative ∆δRS for H(3). Figure 3 shows
a bar diagram with those values and signs.
(B) A new NMR parameter |∆(∆δRS)|, which compares
the shielding supported by H(2) and H(3), makes it possible
to carry out an effective discrimination between the syn
A/anti D and syn B/anti C pairs and therefore to attain an
unambiguous correlation between NMR and absolute ster-
eochemistry.
Figure 1. (a) Four stereoisomers of 1,2,3-prim,sec,sec-triols. (b)
Main shielding effects in the tris-(R)-MPA ester of a syn type A
triol and (d) an anti type C triol. (c) Main shielding effects in the
tris-(S)-MPA ester of a syn type A triol and (e) an anti type C
triol.
the 1H NMR spectra of the corresponding tris-(R)- and tris-
(S)-MPA ester derivatives, easily prepared in one pot by
reaction of the auxiliary and the triol.4
This parameter arises from detailed analysis of the effects
caused by the three MPA units in the syn and anti series. It
shows two clearly different situations, summarized as fol-
lows:
Theoretical and experimental studies (calculations,5 dy-
namic and low-temperature NMR, CD) of the tris-(R)- and
tris-(S)-2-methoxy-2-phenylacetic acid (MPA) esters of syn
and anti triols showed us the way the chemical shifts of H(2)
and H(3) result from the combined action of the shielding
due to the three auxiliaries. This is graphically illustrated in
Figure 1 and can be summarized as follows: (a) In the (R)-
derivative of a syn type A isomer (Figure 1b), H(2) and H(3)
are subjected to neither the shielding from the primary MPA
nor that from the secondary MPAs.6a (b) In the (S)-MPA
derivative of a syn type A isomer (Figure 1c), the shielding
generated by the primary MPA reinforces those produced
by the two secondary MPAs.6b (c) In the (R)-MPA derivative
of an anti type C isomer (Figure 1d), the primary MPA
shields H(3), adding its effect to that caused by the MPA at
C(2), and also shields H(2).6c (d) In the (S)-MPA derivative
of an anti type C triol (Figure 1e), the major shielding
(1) In the syn series, protons H(2) and H(3) are either
unaffected or similarly affected by the shielding caused by
the Ph groups (in tris-(R)- and tris-(S)-derivatives; Figures
1b,c and 2Sf,g) and accordingly the difference between the
∆δRS values for H(3) and for H(2) is expected to be small.
(2) In the anti series, there is a significant difference
between the overall shielding experienced by H(2) and by
H(3) (in tris-(R)- and tris-(S)-derivatives; Figures 1d,e and
2Sh,i): H(2) is heavily shielded in one derivative and H(3)
is heavily shielded in the other. Consequently, the difference
between the ∆δRS values for H(3) and for H(2) is expected
to be high.
Experimental corroboration of that reasoning is presented
in Figure 3. A bar diagram shows the magnitude of the
difference (as absolute value) between the ∆δRS of H(2) (with
its sign) minus the ∆δRS of H(3) (with its sign), expressed
as |∆δRS
- ∆δRSH(3)| ) |∆(∆δRS)|, for triols 1-23.
H(2)
(3) It is present in sugars, itols, and many compounds of pharmaceutical
and biological interest such as Zanamavir (Relenza), sialic acid derivatives,
polyoxamic acid (component of polyoximes), and sphingofungin compo-
nents (see Figure 1S and references in the Supporting Information).
(4) See the Supporting Information for the complete experimental
procedure.
(5) Theoretical calculations [energy minimization by semiempirical
(AM1), and DFT (B3LYP)] were performed with Gaussian 98.
(6) (a) Exactly the same pattern of shielding is expected in the
(S)-derivative of a syn type B isomer (Figure 2S(g), Supporting Information),
(b) for the (R)-MPA derivative of a syn type B isomer (Figure 2S(f),
Supporting Information), (c) for the (S)-MPA derivative of an anti type D
isomer (Figure 2S(i), Supporting Information), and (d) for the (R)-MPA
derivative of the anti type D isomer (Figure 2S(h), Supporting Information).
As predicted, the differences |∆(∆δRS)| are clearly smaller
for the syn than for the anti series. The values experimentally
obtained for the syn A and syn B triols tested (1-12) range
from 0.00 to a maximum of 0.05 ppm, while in the anti
series, the |∆(∆δRS)| values are larger and range from 0.16
up to 0.60 ppm for the triols tested (13-23). These |∆(∆δRS)|
values are so different they allow the discrimination between
(7) The studies show that the diagnostic value of H(2) for this purpose
presents limitations.
4450
Org. Lett., Vol. 8, No. 20, 2006