Synthesis and analysis of the all-(S) side chain of phosphomycoketides: A test of NMR predictions for saturated oligoisoprenoid stereoisomers

Article abstract of DOI:10.1021/jo4005298

(4S,8S,12S,16S,20S)-Pentamethylheptacosan-1-ol has been synthesized and analyzed by resolution-enhanced NMR spectroscopy with the aid of a recent set predicted spectra of all its stereoisomers. The configuration was confirmed, but isomer purity of the sample (～70%) was lower than expected. A truncated analogue, (2S,6S,10S,14S)-2,6,10,14-tetramethylhenicosan-1-ol TBDPS ether, was prepared from a late stage synthetic intermediate. Analysis of its spectra confirmed the configuration and showed that the sample was isomerically pure. The results suggest that a late-stage epimerization, not a failure of an asymmetric synthesis step, caused the formation of minor stereoisomers in the sample of pentamethylheptacosan-1-ol. The study shows the value of the predicted set of oligoisoprenoid spectra and further extends the predictive model to a new subclass of compounds.

Full text of DOI:10.1021/jo4005298

Article
pubs.acs.org/joc
Synthesis and Analysis of the All‑(S) Side Chain of
Phosphomycoketides: A Test of NMR Predictions for Saturated
Oligoisoprenoid Stereoisomers
Jeffrey Buter,† Edmund A.-H. Yeh,‡ Owen W. Budavich,‡ Krishnan Damodaran,‡ Adriaan J. Minnaard,*,†
and Dennis P. Curran*,‡
†Stratingh Institute for Chemistry, University of Groningen, Nijenborgh 7, 9747 AG Groningen, The Netherlands
‡Department of Chemistry, University of Pittsburgh, Pittsburgh, Pennsylvania 15260, United States
S
* Supporting Information
ABSTRACT: (4S,8S,12S,16S,20S)-Pentamethylheptacosan-1-ol has
been synthesized and analyzed by resolution-enhanced NMR
spectroscopy with the aid of a recent set predicted spectra of all its
stereoisomers. The configuration was confirmed, but isomer purity of
the sample (∼70%) was lower than expected. A truncated analogue,
(2S,6S,10S,14S)-2,6,10,14-tetramethylhenicosan-1-ol TBDPS ether,
was prepared from a late stage synthetic intermediate. Analysis of
its spectra confirmed the configuration and showed that the sample was isomerically pure. The results suggest that a late-stage
epimerization, not a failure of an asymmetric synthesis step, caused the formation of minor stereoisomers in the sample of
pentamethylheptacosan-1-ol. The study shows the value of the predicted set of oligoisoprenoid spectra and further extends the
predictive model to a new subclass of compounds.
mannose and a long, aliphatic chain linked together as a
phosphate diester. In MPM biosynthesis, the polyketide side
chain is formed from a fatty acid precursor by repetitive
incorporation of malonyl and methyl malonyl CoA followed by
full reduction.
INTRODUCTION
■
β-Mannosyl phosphomycoketides (MPMs) 1 and 2 (Figure 1a)
are potent T-cell antigens produced by Mycobacterium avium
and Mycobacterium tuberculosis.1 They are composed of β-D-
Although MPM’s side chains have a polyketide origin, a large
part of the chain 3c (C2−C21) bears methyl groups on every
fourth carbon and therefore resembles a saturated oligoiso-
prene. In polymer terminology, such repeat units are also called
alternating propylene/ethylene co-oligomers or 1-methyltetra-
methylene oligomers.2 Rigorously assigning the configurations
of the five side chain stereocenters of such oligomers is difficult.
Crich intentionally synthesized a mixture of MPMs 1 (R = n-
C4H9) with complete stereocontrol of the mannosyl anomeric
center but with all possible configurations of the side-chain
stereocenters.3 Analysis of this mixture of 32 stereoisomers
confirmed the constitution of 1. Minnaard later synthesized a
pentamethylheptacosan-1-ol side chain with all five side-chain
stereocenters in the (S)-configuration (all-(S)-3c, Figure 1b)
and coupled this to β-mannose phosphate to make MPM 2 (R
= n-C6H13).4 This sample was assayed against Crich’s mixture
and a natural product sample and exhibited activity similar to
that of the natural product.5
The five stereocenters of the side chains of 1 and 2 are
biosynthesized in an iterative way by the polyketide synthase
Pks12 and therefore presumably have the same configuration.1,6
This means that the 1,4-relative configurations of the methyl-
Received: March 13, 2013
Published: April 12, 2013
Figure 1. Structures of MPM natural products and side chains.
© 2013 American Chemical Society
4913
dx.doi.org/10.1021/jo4005298 | J. Org. Chem. 2013, 78, 4913−4918

The Journal of Organic Chemistry
Article
a
Scheme 1. Synthesis of Two Samples for NMR Analysis
a
3c is the complete MPM side chain, and 7 is a saturated derivative of key late-stage intermediate 4.
branched stereocenters are syn. Of the two remaining
configurations, all stereocenters (R) or all (S), the results of
the bioassays support the all-(S) assignment of 2.4a,5 In
addition, the crystal structure of CD1c bound to all-(S)-MPM 2
has been reported recently.7 However, assignment of the
absolute configuration as all-(S) contradicts a predictive model
by Leadley and co-workers.8 Therefore, spectroscopic or
chemical means to assess the side-chain configurations is
needed both to confirm the structure of the natural product and
to assay structures and purities of synthetic samples.9
conjugate additions reactions in with high levels of stereo-
selectivity (>95%).
The two compounds analyzed in detail were late-stage
derivative 7, which is a phytanyl-type oligoisoprenoid with a
silyl-protected hydroxy group on the left end and four repeating
stereocenters, and 3c, which is the full MPM side chain with
five repeating stereocenters. Both samples were derived from
unsaturated silyl ether 4, itself made as a 2/1 mixture of E/Z
isomers in a Kocienski−Julia reaction.12
The silyl ether 4 was desilylated (95%), and the resulting
alcohol was oxidized to an aldehyde (81%). This in turn was
added to the anion derived from sulfonyl tetrazole 5 to give 6 as
an E/Z mixture at both the existing (C14) and the newly
formed (C6) alkene (81%). Hydrogenation of the alkenes in 6
occurred simultaneously with reductive debenzylation to
provide the MPM side chain sample 3c.
Recently we synthesized all four diastereomers of the
truncated MPM side-chain model 4,8,12-trimethylnonadecanol
3a (Figure 1c).10 The methyl regions of both the H and 13C
1
NMR spectra of the four isomers exhibited small but reliable
differences depending on whether the nearest neighbor methyl
groups were syn or anti. We used the data to predict the 1H and
13C NMR spectra of all stereoisomers of higher saturated
oligoisoprenoids including 3b (8 isomers) and 3c (16 isomers).
Unfortunately, it is not possible to test these predictions
retrospectively with published spectra because special con-
ditions for processing are required to enhance resolution.
Here we report the synthesis of a new sample of all-(S)-3c.
Analyses of resolution-enhanced NMR spectra prove that the
sample is predominately all-(S) but also show that the isomer
ratio is lower than expected. Further analysis of a lower
homologue prepared from synthetic intermediates suggests that
the impurities arose not because an asymmetric synthesis step
was compromised, but instead because of a late-stage partial
epimerization. Just as the predicted spectra help to analyze the
experimental spectra, the reverse is also true. The new chemical
shift values obtained from the experimental spectra help to
refine and expand the NMR model.
Subsequent NMR studies of the final product 3c (see below)
led us to question the stereochemical integrity of 4. To assess
this, we saturated a small sample of the E/Z isomer mixture of 4
to give 7. This sample was isomerically pure (see below) so by
implication all the reactions leading to 4 (see Scheme S1 in the
Supporting Information) occurred with high stereoselectivity,
and the so-formed stereocenters were retained with fidelity.
Analysis of the Spectra of 3c. To assess the structure and
isomeric purity of 3c, the observed resonances in the methyl
1
regions of both its H and 13C NMR spectra were compared
with recent predictions.10 The basis for predicting the
resonances of any isomer of 3c (n = 3) from the observed
resonances of 3a (n = 1) is shown in Figure 2. Briefly, we divide
the spectrum into three parts, the left-end, the middle, and the
right-end, and then add in resonances of the branched methyl
groups from the model set (all isomers of 3a) that have the
appropriate stereochemical relationship to match those in 3c.
End Me groups can be syn or anti, while middle ones can be
syn/syn, anti/anti, or anti/syn (or syn/anti).
RESULTS AND DISCUSSION
■
Synthesis of the Samples. The synthesis of the new
sample of 3c was based on the catalytic enantioselective
synthesis of β-D-mannosyl phosphomycoketide reported by
Minnaard and co-workers.4a,11 The late stages of this synthesis
are summarized in Scheme 1, while the Supporting Information
provides complete information on reagents and yields in
Schemes S1 and S2. All five of the stereocenters in penultimate
intermediates 4 and 5 were synthesized by asymmetric
To predict a spectrum of the all-syn isomer of 3c (all-(S)-3c
or all-(R)-3c), we simply take the spectrum of all-syn 3a and
multiply the middle resonances by three (because all three
middle Me groups have syn/syn relationships), as shown in
Figure 2a. Figure 2b lists the remaining resonances from the
complete model set. These are not used for all-syn-3c but are
needed to predict spectra of its stereoisomers.
4914
dx.doi.org/10.1021/jo4005298 | J. Org. Chem. 2013, 78, 4913−4918

The Journal of Organic Chemistry
Article
Figure 2. Basis for predictions of NMR spectra of all-syn-3c and
stereoisomers.
Standard spectra of all-syn-3c were first recorded in the usual
way at 400 MHz. These compared favorably to the spectra of
the prior sample4a but as expected did not provide information
about isomer identity or purity. A new set of spectra were then
1
recorded at 700 MHz for H and 175 MHz for 13C, and the
data sets were processed by the Traficante algorithm13 for
resolution enhancement.
1
Figure 3a,b shows the methyl regions of the 13C and H
NMR spectra, both predicted (bottom spectrum) and actual
(top spectrum). The processing renders the spectra of such
samples interpretable for the first time.
The predicted 13C NMR spectrum of all-syn-3c (Figure 3a,
bottom) shows three peaks: the left-end Me resonates at 19.67
ppm, while the right-end Me resonates at 19.77 ppm. The
predicted chemical shift of the three middle methyl groups (on
C8, C12, C16) is 19.79 ppm, with a relative intensity of three.
The experimentally obtained 13C NMR spectrum (Figure 3a,
top) is at first glance much more complex. There are 10
resonances grouped in two sets of five with a ratio of about 70/
30. Thus, the sample is not isomerically pure. Three of the five
peaks of the major set (19.79, 19.77, and 19.67 ppm) match
perfectly (<10 ppb difference) to the predicted peaks. The two
most downfield major peaks at 19.80 and 19.81 ppm are slightly
downfield from their predicted location at 19.79 ppm.
Figure 3. Comparison of the predicted and the experimental NMR
spectra of the branched methyl group region for all-syn-3c.
terminal methyl group (C29 on the far right end). This triplet
has the same predicted chemical shift in all of the isomers, so its
value is not diagnostic. This leaves three major doublets in a
ratio of 1:3:1 for the five branched Me groups of the major
isomers. The smaller ones at 0.874 and 0.840 ppm are the left-
and right-end methyl groups. As in the predicted 13C NMR
spectrum, the three middle methyl groups overlap, now at
0.844 ppm. The major resonances of the spectrum match the
predictions of the all-syn isomer closely. Again, the match is
clearly the best among all 16 predicted spectra.10
We compared the major set of resonances in the
experimental spectrum of all-syn-3c to the predicted spectra
of all 16 isomers of 3c,10 and none matches as well as the
spectrum of the all-syn-isomer. Thus, we conclude that the five
major peaks in the actual spectrum in Figure 3a belong to all-
syn-3c. This means that the predicted spectrum in Figure 3a is
only partially correct. Because the model 3a is too simple (it
has only one middle unit), it cannot anticipate the very small
differences in chemical shifts of the three “repeating” methyl
groups in the middle of all-syn-3c. Indeed, given their chemical
and stereochemical similarity, it is remarkable that separate
resonances for these methyl groups are observed.
1
The experimental H NMR spectrum again has a minor set
of resonances; at least three minor doublets can be seen in the
1D spectrum. To find the other two minor resonances and to
1
correlate all the resonances, we conducted a H 13C COSY
experiment. Figure 4 shows two slices of the resulting
spectrum: plot (a) is a higher slice that shows only the
crosspeaks of the five major resonances, while plot (b) is a
lower slice that shows both sets, but only the five minor
crosspeaks are highlighted.
1
The methyl regions of the predicted and experimental H
NMR spectra of all-syn-3c are shown in Figure 3b. The
predicted spectrum is again simpler than the actual one. Now
there is an additional resonance; the triplet at 0.882 ppm is the
4915
dx.doi.org/10.1021/jo4005298 | J. Org. Chem. 2013, 78, 4913−4918

The Journal of Organic Chemistry
Article
two resonances coincide, while the third is a smidge downfield
(∼1 ppb), appearing as a shoulder on the other two.
In contrast to the clear situation with the major set of five
resonances, assignment of the minor set of five resonances is
not straightforward. Four of the 13C and 1H resonances
correspond well to other resonances predicted by the model
(see unused resonances in Figure 2b): 19.73/0.842 ppm is a
middle methyl group with a syn/anti (or anti/syn relationship)
to its neighbors; 19.70/0.839 ppm is the right methyl group
with an anti neighbor; 19.66/0.841 ppm is a middle methyl
group with an anti/anti relationship; and 19.61/0.872 ppm is
left methyl group with an anti neighbor.
There is also a minor peak at 19.68 ppm in the 13C NMR
that is not in the model set. This peak is only 10 ppb downfield
from model peak 19.67, a left methyl group with a syn neighbor.
However, if this assignment were correct, then that 13C peak
1
should correlate to a H resonance at 0.874 ppm. Instead, it
correlates to a resonance at 0.842 ppm.
In short, the major isomer in the new sample of 3c is clearly
the all-syn isomer. However, because of the simplicity of the
model (it provides only seven resonances for all peaks in all
possible isomers), we cannot yet assign the minor peaks. If they
come from a single isomer, then one of the peaks (19.68 ppm)
is poorly predicted in the 13C NMR spectrum. If they come
from a mixture of isomers, then there must be other minor
peaks under the major peaks in varying ratios, so direct
interpretation is not easy.
Analysis of the Spectra of 7. We next conducted a series
of experiments trying to localize the stage of the synthesis
where the minor isomer(s) originated. Small samples of key
intermediates saved during the synthesis, especially the sample
of 4 (Scheme 1), proved very helpful. This alkene could not be
analyzed directly for stereochemical integrity because it already
contains E/Z isomers from the Julia−Kocienski reaction. So
about 2 mg of 4 was carefully hydrogenated (Pt, H2, CH2Cl2/
MeOH) to make a sample of 7 whose spectra were recorded as
above.
The predicted spectrum of all-syn-3b was used for
comparison, and the structures of 3b and 7 are compared in
Figure 5. Notice that while the right-end and middle fragments
of 7 and all-syn-3b are the same, the left-end fragment is
different. Sample 7 has two carbon atoms fewer than model 3b
at the left end, and its hydroxy group is protected with a
1
Figure 4. Branched methyl region of the H 13C COSY of all-syn 3c.
All of the major peaks correlate as predicted by the model in
Figure 2a, thus reinforcing the assignment of the major
component of 3c as all-syn. The reason that the largest doublet
1
at 0.844 ppm in the H NMR spectrum also appears to be
further resolved into three peaks is revealed. The right shoulder
comes from a minor isomer, correlated at 19.73 ppm in the 13
C
spectrum, while the left shoulder comes from one of the middle
carbons of the major isomer correlated at 19.81 ppm. The large
center part correlates to the other two middle methyl groups in
1
the major isomer at 19.80 and 19.79 ppm. So in the H NMR
Figure 5. Predicted spectra of all-syn-3b are used to interpret the
actual spectra of 7. Only the middle and right-end predictions can be
used because the left ends are different.
spectrum, the prediction that three middle methyl groups of all-
syn-3c would coincide is again wrong, but only just. This time,
4916
dx.doi.org/10.1021/jo4005298 | J. Org. Chem. 2013, 78, 4913−4918

The Journal of Organic Chemistry
Article
TBDPS (tert-butyldiphenylsilyl) ether. Clearly the predicted
left-end resonances for all-syn-3b cannot be applied to the left
end of 7.
of resonances has no equivalent in the model, but the values are
not crucial to the stereochemical assignment because the
chemical shifts of the next methyl group over (Me23) also
contain the information about whether the relationship of the
two is syn or anti.
The methyl region of the experimental 1H NMR spectrum of
7 is shown in the top of Figure 6. There is only one set of
Now compare the predicted and actual resonances for the
remaining Me groups (Me23, Me24, and Me25) in structure 7,
starting at the right end of the molecule. In both the 1H and the
13C NMR spectra, the actual and predicted resonances for the
right-end methyl group (Me25) and its neighbor (Me24) are
spot on or almost spot on. At 0.830 ppm (1H) and 19.76 (13C)
ppm, the actual resonances of the next methyl group in (Me23)
are slightly different from both Me24 and the prediction. As
above, this is again a deficiency of the model, which provides
only a single middle resonance that is used to model both Me23
and Me24. It is sensible that the model is spot on for Me24,
which is further away from the left end (the part that differs
between the model and the actual sample), and slightly off for
Me23, which is closer to the left end.
1
The values of the experimental H NMR spectra of 7 in
Table 1 are much closer to the predicted values for the all-syn
isomer of 3b than to any of the other seven isomers.10 Thus,
both the relative configuration of the sample (all-syn) and the
high level of purity (≥95%) are confirmed.
These experiments show that the four stereocenters of key
synthetic intermediate 4 are intact and have the configurations
predicted by the various asymmetric reactions used to
introduce them (Scheme 1). Thus, the minor stereoisomer(s)
in sample 3c must have formed in a late stage of the synthesis,
during or after the coupling of the last two fragments 4 and 5
(Scheme 1). Primary candidate steps for the epimerization are
the Julia−Kocienski reaction (the aldehyde reactant has an
adjacent stereocenter) and the final hydrogenation reaction
(alkene migration to form a trisubstituted double bond, prior to
hydrogenation, results in loss of stereointegrity14).
Finally, this exercise establishes a new left-end model
compound with a CH2OTBDPS group from which spectra of
saturated isoprenoids and related reduced polyketides can be
predicted. Compounds such as 7 terminating in CH2OTBDPS
on the left end will have a chemical shift of 0.830 and 19.76
ppm for the left-end methyl group (Me22) if it is syn to the next
methyl group (Me23). On the basis of the established trends
with 3a, we predict that Me22 will resonance a little further
upfield in both the 1H and 13C NMR spectra if it is anti to Me23
(about 2 and 60 ppb, respectively, see Figure 2). From the
chemical shifts in Table 1, small upfield adjustments also need
to be made for the resonances of Me23 in the syn/syn isomer to
predict the spectra of isomers with this group syn/anti (or anti/
syn) and anti/anti to Me22/Me24. The magnitudes of the
adjustments can again be estimated from the values for the
analogous middle methyl groups in Figure 2.
1
Figure 6. Expansions of the methyl regions of the H (top) and 13C
1
(bottom) NMR spectra of 7. The H NMR expansion contains the
terminal methyl group (C21) resonance, but the 13C spectrum does
not.
resonances: four doublets and one triplet (the terminal methyl
group on the end of the chain, C21). The resonances were
assigned by 2D NMR experiments (see Supporting Informa-
tion), and these assignments are collected in Table 1. Likewise,
the 13C NMR spectrum showed one set of resonances (bottom
of Figure 6 and Table 1). Clearly this sample is a single isomer.
To assign the configuration of 7, we compared the predicted
methyl resonances of all-syn 7 with the actual resonances, as
summarized in Table 1. The observed resonances for left-end
methyl group (Me22) appear at 0.917 ppm in the H NMR
1
spectrum and 16.99 ppm in the 13C NMR spectrum. This pair
Table 1. Comparison of Methyl Resonances of the Actual
Spectra of 7 with Predicted Resonances for the All-syn
Isomer of 7 Derived from the Predicted Spectra of all-syn-
3b; Data in ppm
1H
13C
These new resonances can be added to the existing middle
(used for Me24) and right (used for Me25) resonances in Figure
2 to fill out predicted spectra for any isomer of 7 or a higher
oligomer. Because the model 7 has four resonances, every
isomer of 7 (or a higher oligomer) will also have four predicted
resonances. This is an improvement over the use of 3a to
model 3b and 3c, where two or three middle Me groups have
the same predicted chemical shift in some isomers, even though
in reality there are small differences.
resonance
obsd
pred
obsd
pred
Me22
Me23
Me24
Me25
0.917
0.830
0.845
0.840
a
16.99
19.76
19.80
19.76
a
b
b
0.844
0.844
0.840
19.79
19.79
19.77
b
c
b
c
a
b
The model does not predict the left-end resonance. Middle
resonance, see Figure 2. Right-end resonance, see Figure 2.
c
4917
dx.doi.org/10.1021/jo4005298 | J. Org. Chem. 2013, 78, 4913−4918

The Journal of Organic Chemistry
Article
CONCLUSIONS
ASSOCIATED CONTENT
* Supporting Information
Copies of NMR spectra of 3c, 4, and 7. This material is
■
■
S
In summary, for the first time it has been possible through
NMR spectroscopy both to assess isomer purity and to assign
relative configurations of saturated oligoisoprenoids such as 3c.
available free of charge via the Internet at http://pubs.acs.org.
Recent predictions of the methyl regions of the H and 13C
1
AUTHOR INFORMATION
Corresponding Author
*E-mail: a.j.minnaard@rug.nl; curran@pitt.edu.
Notes
■
NMR spectra of all 16 stereoisomers of 3c enabled this
analysis.10 The predicted H NMR spectrum of all-syn-3c is
1
very close to the actual spectrum, but the predicted 13C NMR
spectrum proved to be too simple in one aspect: the three
resonances from the middle methyl groups were predicted to
coincide because they all have the syn/syn relationship, but they
did not. However, the observed differences between predicted
and actual spectra were small compared to the differences
expected for other stereoisomers, so the value of the 13C NMR
predictions was not compromised.
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
The Groningen group thanks NWO-CW for a VICI grant to
A.J.M. The Pittsburgh group thanks the National Institutes of
Health for funding this work and for funds to help purchase a
700 MHz NMR spectrometer.
The analysis of the synthetic sample of 3c showed that the
all-(S) (all-syn) isomer was indeed the major component,
present to the extent of about 70%. Unexpectedly, a minor
stereoisomer component (or components) was present to the
extent of about 30%. We cannot yet identify the minor
component or pinpoint where it was introduced. However, by
similar comparison of actual and predicted spectra, we could
show that 7 (derived from a key synthetic intermediate 4
bearing four of the five stereocenters) is both pure (>90%) and
has the expected all-(S) configuration. This narrows the source
of the problem to the last few steps of the synthesis.
Perhaps most importantly, the values from the predicted
spectra of compounds like 3c can now be leveraged to related
compounds bearing different right or left ends. As an example,
intermediate 7 with four stereocenters has a left end different
from that of 3a−c. Even though we only made one of the eight
possible stereoisomers of 7, we can now combine the new data
obtained for 7 with the data for 3a to assemble predicted
spectra of the other seven isomers of 7 and its higher oligomers.
This ability to directly analyze complex, saturated oligoisopre-
noids is a powerful tool to clarify a heretofore cloudy situation
with respect to stereoisomer structure and purity of synthetic
and natural samples.
REFERENCES
■
(1) Moody, D. B.; Ulrichs, T.; Muhlecker, W.; Young, D. C.; Gurcha,
̈
S. S.; Grant, E.; Rosat, J.-P.; Brenner, M. B.; Costello, C. E.; Besra, G.
S.; Porcelli, S. A. Nature 2000, 404, 884−888.
(2) Zetta, L.; Gatti, G.; Audisio, G. Macromolecules 1978, 11, 763−
766.
(3) Crich, D.; Dudkin, V. J. Am. Chem. Soc. 2002, 124, 2263−2266.
(4) (a) van Summeren, R. P.; Moody, D. B.; Feringa, B. L.;
Minnaard, A. J. J. Am. Chem. Soc. 2006, 128, 4546−4547. (b) Ferrer,
C.; Fodran, P.; Barroso, S.; Gibson, R.; Hopmans, E. C.; Sinninghe
́
Damste, J.; Schouten, S.; Minnaard, A. J. Org. Biomol. Chem. 2013, 15,
2482−2492.
(5) de Jong, A.; Arce, E. C.; Cheng, T.-Y.; van Summeren, R. P.;
Feringa, B. L.; Dudkin, V.; Crich, D.; Matsunaga, I.; Minnaard, A. J.;
Moody, D. B. Chem. Biol. 2007, 14, 1232−1242.
(6) Matsunaga, I.; Bhatt, A.; Young, D. C.; Cheng, T.-Y.; Eyles, S. J.;
Besra, G. S.; Briken, V.; Porcelli, S. A.; Costello, C. E.; Jacobs, W. R.;
Moody, D. B. J. Exp. Med. 2004, 200, 1559−1569.
́
(7) Scharf, L.; Li, N.-S.; Hawk, A. J.; Garzon, D.; Zhang, T.; Fox, L.
M.; Kazen, A. R.; Shah, S.; Haddadian, E. J.; Gumperz, J. E.;
Saghatelian, A.; Faraldo-Gomez, J. D.; Meredith, S. C.; Piccirilli, J. A.;
́
Adams, E. J. Immunity 2010, 33, 853−862.
(8) Kwan, D. H.; Sun, Y.; Schulz, F.; Hong, H.; Popovic, B.; Sim-
Stark, J. C. C.; Haydock, S. F.; Leadlay, P. F. Chem. Biol. 2008, 15,
1231−1240.
(9) Ohrui, H. Proc. Jpn. Acad., Ser. B 2007, 83, 127−135.
(10) Yeh, E. A.; Kumli, E.; Damodaran, K.; Curran, D. P. J. Am.
Chem. Soc. 2013, 135, 1577−1584.
EXPERIMENTAL SECTION
■
(11) (a) van Summeren, R. P.; Reijmer, S. J. W.; Feringa, B. L.;
Minnaard, A. J. Chem. Commun. 2005, 1387−1389. (b) Naito, J.;
Kuwahara, S.; Watanabe, M.; Decatur, J.; Bos, P. H.; Van Summeren,
R. P.; Horst, B. T.; Feringa, B. L.; Minnaard, A. J.; Harada, N. Chirality
2008, 20, 1053−1065.
The synthesis of intermediates 3c and 6 followed published routes.
Experimental details and characterization data can be found in refs 4
and 11. Detail reaction conditions and yields are shown in the
Supporting Information, Schemes S1 and S2.
tert-Butyldiphenyl(((2S,6S,10S,14S)-2,6,10,14-tetramethyl-
henicosyl)oxy)silane (7). Dry 10 wt % Pt/C (0.003 g, 0.002 mmol)
was added to 6 (0.003 g, 0.005 mmol) in MeOH/CH2Cl2 (3:1, 1 mL).
The suspension was flushed three times with H2 (three vacuum/H2
cycles) and stirred under an atmosphere of H2 (1 atm, balloon) at rt
for 24 h. The reaction mixture was filtered through a silica gel plug and
(12) Blakemore, P. R. J. Chem. Soc., Perkin Trans. 1 2002, 2563−
2585.
(13) Traficante, D. D.; Nemeth, G. A. J. Magn. Reson. 1987, 71, 237−
245.
(14) (a) Dandapani, S.; Jeske, M.; Curran, D. P. J. Org. Chem. 2005,
70, 9447−9462. (b) Teichert, J. F.; den Hartog, T.; Hanstein, M.;
Smit, C.; ter Horst, B.; Hernandez-Olmos, V.; Feringa, B. L.;
Minnaard, A. J. ACS Catal. 2011, 1, 309−315.
1
concentrated to yield the hydrogenation product 7 quantitatively. H
NMR (CDCl3, 700 MHz, ppm) δ 7.66 (d, J = 6.8 Hz, 4 H), 7.41 (t, J =
7.3 Hz, 2 H), 7.37 (t, J = 7.2 Hz, 4 H), 3.51 (dd, J = 5.6, 9.8 Hz, 1 H),
3.43 (dd, J = 6.4, 9.8 Hz, 1 H), 2.22 (t, J = 7.6 Hz, 1 H), 1.63 (m, 3 H),
1.13−1.42 (m, 30 H), 1.05 (s, 9 H), 0.92 (d, J = 6.7 Hz, 3 H), 0.88 (t, J
= 7.0 Hz, 3 H), 0.82−0.85 (m, 9 H); 13C NMR (CDCl3, 175 MHz,
ppm) δ 135.61, 134.13, 134.11, 129.43, 127.52, 68.88, 37.42, 37.40,
37.38, 37.34, 37.05, 35.71, 33.46, 32.80, 32.76, 32.75, 31.92, 29.99,
29.69, 29.66, 29.65, 29.64, 29.60, 29.40, 29.36, 27.08, 26.85, 24.47,
24.45, 24.38, 22.69, 19.79, 19.75, 19.74, 19.30, 16.97, 14.13.
4918
dx.doi.org/10.1021/jo4005298 | J. Org. Chem. 2013, 78, 4913−4918