Stereoselective synthesis of anti-N-protected 3-amino-1,2-epoxides by nucleophilic addition to N-tert-butanesulfinyl imine of a glyceraldehyde synthon

Article abstract of DOI:10.1021/jo900643b

(Chemical Equation Presented) A di-O-TBS protected glyceraldehyde synthon was condensed with Ellman's reagent to form a bench-stable N-tert-butanesulfinyl imine 6, which served as a common intermediate for the stereoselective introduction of various R groups. The Ellman adducts were converted to useful multifunctional intermediates 18a-i in one pot. The alcohols 18a-i were efficiently elaborated to both known and novel anti-N-protected-3-amino-1,2- epoxides in two steps. Compound 2a is a key intermediate toward HIV protease inhibitors.

Full text of DOI:10.1021/jo900643b

pubs.acs.org/joc
Stereoselective Synthesis of anti-N-Protected 3-Amino-1,2-epoxides by
Nucleophilic Addition to N-tert-Butanesulfinyl Imine of a Glyceraldehyde
Synthon†
Scott S. Harried,* Michael D. Croghan, Matthew R. Kaller, Patricia Lopez, Wenge Zhong,
Randall Hungate, and Paul J. Reider‡
Department of Medicinal Chemistry, Amgen Inc., Thousand Oaks, California 91320-1799.
‡Current address: Dept. of Chemistry, Princeton University, Princeton, NJ 08544.
sharried@amgen.com
Received April 3, 2009
A di-O-TBS protected glyceraldehyde synthon was condensed with Ellman’s reagent to form a
bench-stable N-tert-butanesulfinyl imine 6, which served as a common intermediate for the
stereoselective introduction of various R groups. The Ellman adducts were converted to useful
multifunctional intermediates 18a-i in one pot. The alcohols 18a-i were efficiently elaborated to
both known and novel anti-N-protected-3-amino-1,2-epoxides in two steps. Compound 2a is a key
intermediate toward HIV protease inhibitors.
Introduction
introduction of diversity limits both the speed and efficiency
for the synthesis of new analogues.
The hydroxyethylamine (HEA) isostere is a common
structural motif present in several approved HIV protease
inhibitors, such as Amprenavir 1 (Scheme 1).1 anti-N-BOC-
3-Amino-1,2-epoxides are commonly used as key intermedi-
ates toward these important drugs.2 Many of the published
syntheses of these N-BOC-amino epoxides rely on either N-
BOC protected amino acids or their derivatives. These routes
are very efficient and cost-effective for the synthesis of a
single targeted epoxide as an intermediate toward an active
pharmaceutical ingredient. However, from a medicinal
chemistry perspective, these amino-acid-based routes are
hampered by a few shortcomings. The main disadvantage
of these routes is the early introduction of P1 molecular
diversity.3 The P1 group is introduced at the first step of
the synthetic sequence when an amino acid or its derivative
is used to synthesize a 3-amino-1,2-epoxide. The early
Another shortcoming of these amino-acid-based routes is
the potential for racemization of R chiral esters or activated
acid intermediates.4 Additionally, several of these routes use
potentially hazardous reagents, such as diazomethane.5 Che-
mists2 have also utilized N,N-dibenzyl amino aldehydes6 (4 in
Scheme 1) in order to achieve the desired 2,3-anti relationship
of the N-protected amino epoxides. Again, with this strategy,
the P1 diversity is introduced at the first step of the synthetic
sequence and is usually applied to amino acid inputs.
Results and Discussion
The powerful, flexible, and practical sulfinyl imine chemis-
try pioneered by Davis7 and extended by Ellman8 and others9
(4) Wang, D.; Schwinden, M., D.; Radesca, L.; Patel, B.; Kronenthal, D.;
Huang, M.; Nugent, W., A. J. Org. Chem. 2004, 69, 1629–1633.
(5) Albeck, A.; Persky, R. Tetrahedron 1994, 50, 6333–6346.
(6) Reetz, M. T. Angew. Chem., Int. Ed. Engl. 1991, 30, 1531–1546.
(7) Davis, F. A. J. Org. Chem. 2006, 71, 8993–9003.
† This manuscript is dedicated to Professor Yoshito Kishi.
(1) Bursavich, M. G.; Rich, D. H. J. Med. Chem. 2002, 45, 541–558.
(2) (a) Izawa, K.; Onishi, T. Chem. Rev. 2006, 106, 2811–2827 and
references therein. (b) Ghosh, A. K.; Bilcer, G.; Schiltz, G. Synthesis 2001,
2203–2229 and references therein .
(8) (a) Ellman, J. A.; Owens, T. D.; Tang, T. P. and references therein.
Acc. Chem. Res. 2002, 35, 984–995. (b) Senanayake, C. H.; Krishnamurthy,
D.; Lu, Z.; Han, Z.; Gallou, I. Aldrichimica Acta 2005, 38, 93–104 and
references therein .
(3) For a Perspective on protease inhibitors and standard nomenclature,
see: Leung, D.;Abbenante, G.; Fairlie, D., P. J. Med. Chem. 2000, 43, 305–341.
(9) Ferreira, F.; Botuha, C.; Chemla, F.; Perez-Luna, A. Chem. Soc. Rev.
2009, 38, 1162–1186.
DOI: 10.1021/jo900643b
r
Published on Web 07/08/2009
J. Org. Chem. 2009, 74, 5975–5982 5975
2009 American Chemical Society

Harried et al.
JOCArticle
SCHEME 1. Conventional Disconnection of anti-N-BOC-3-Amino-1,2-epoxides
seemed to be a perfect solution for our medicinal chemistry
focused diversity problem. We decided to frame shift the key
disconnection of the epoxide 2a over to the formation of the
C.3-C.4 carbon-carbon bond instead of the more conven-
tional C.1-C.2 carbon-carbon bond disconnection
(Scheme 2).10 In doing so, we have gained a few key advan-
tages. First, we delayed the introduction of our P1 diversity.
Second, we released ourselves from the necessity of starting
with an amino-acid-based input. Third, we mitigated the pro-
blem of epimerization of a stereocenter R to a carbonyl. Appli-
cation of an Ellman tert-butylsulfinyl imine disconnection to
epoxide 2a gives di-O-TBS protected imine 6 and an organo-
metallic reagent 5, many of which are commercially available.
Our initial synthetic challenge was to produce multigram
quantities of a glyceraldehyde synthon for the formation of
the tert-butylsulfinyl imine 6. Both antipodes of glyceralde-
hyde acetonide or related acetals11 are widely used inter-
mediates in the field of synthetic organic chemistry.12
However, we found the acidic hydrolysis conditions required
to remove the acetonide diol protecting group13 to be
incompatible with the standard acidic hydrolysis conditions
commonly used to remove the chiral sulfinyl group. There-
fore, we decided to synthesize a glyceraldehyde equivalent
that has the diol functional group protected as two silyl
ethers instead of a acetal. Corey and co-workers14 have
shown that allylic 4-methoxybenzoates are excellent sub-
strates for the Sharpless asymmetric dihydroxylation15 reac-
tion (AD) in terms of yield and enantiomeric purity.
The required allylic benzoate is readily prepared from com-
mercially available allyl alcohol and 4-methoxybenzoyl
chloride on a mole scale (Scheme 3). The desired absolute
stereochemistry was established using the commercially
SCHEME 2. tert-Butylsulfinyl Imine Disconnection of 2a
available (DHQ)2PHAL ligand in the Sharpless AD reac-
tion. The solid diol 8 was produced in greater than 85% yield
with a 95-98% ee and then protected with 2.2 equiv of
TBSOTf and TEA in DCM to give di-O-TBS ether 9 in good
yield. The 4-methoxybenzoate was reductively removed with
Dibal-H in DCM to give the di-O-TBS protected primary
alcohol 10. The primary alcohol 10 could be stored for
months on the benchtop without loss of optical purity. The
2,3-di-O-TBS glyceraldehyde synthon 1116 could be pro-
duced in good yield by either Dess-Martin17 or Swern18
oxidation conditions without epimerization of the R-silyloxy
stereocenter or β-elimination.
The condensation of Ellman’s tert-butylsulfinamide re-
agent with the glyceraldehyde synthon 11 and subsequent
nucleophilic addition of organolithium or Grignard reagents
were our next challenges. The Ellman laboratory has formed
the N-tert-butanesulfinyl imines of protected lactals without
epimerization at the R-stereocenter.19 Moreover, they ob-
served excellent stereoselective addition of either EtMgBr or
PhMgBr in THF with TMEDA at -78 °C in good yields.
The anti addition product, which was consistent with both
the Felkin-Ahn20 and Cornforth21 models, was formed in a
(10) For a conceptually similar reaction with N-benzylimine of glycer-
aldehyde acetonide, see: (a) Badorrey, R.; Cativiela, C.; Diaz-de-Villegas, M.
D.; Galvez, J. A. Tetrahedron 1997, 53, 1411–1416. (b) Badorrey, R.;
Cativiela, C.; Diaz-de-Villegas, M. D.; Galvez, J. Tetrahedron 2002, 58,
341–354. (c) For a review on additions of organometallic reagents to imines,
see: Bloch, R. Chem. Rev. 1998, 98, 1407–1438.
(11) (a) Chattopadhyay, A.; Mamdapur, V. R. J. Org. Chem. 1995, 60,
585–587 and references therein . (b) Schmid, C. R.; Bryant, J. D.; Dowlatze-
dah, M.; Phillips, J. L.; Prather, D., E.; Schantz, R. D.; Sear, N. L.; Vianco,
C. S. J. Org. Chem. 1991, 56, 4056–4058.
(12) Jurczak, J.; Pikul, S.; Bauer, T. Tetrahedron 1986, 42, 447–488.
(13) Greene, T. W.; Wuts, P. G., M. Protective Groups in Organic
Synthesis, 3rd ed.; John Wiley & Sons, Inc.: New York, NY, 1999.
(14) Corey, E. J.; Guzman-Perez, A; Noe, M., C. J. Am. Chem. Soc. 1995,
117, 10805–10816.
(16) (a) For the synthesis of 11 from D-mannitol, see: Jurczak, J.; Bauer,
T.; Chmielewski, M. Carbohydr. Res. 1987, 164, 493–498. (b) For a route to
di-OBn glyceraldehyde, see: Schreiber, S. L.; Satake, K. Tetrahedron Lett.
1986, 27, 2575–2578.
(17) Dess, D. B.; Martin, J. C. J. Org. Chem. 1983, 48, 4155–4156.
(18) Mancuso, A. J.; Swern, D. Synthesis 1981, 165–185.
(19) (a) Evans, J. W.; Ellman, J. A. J. Org. Chem. 2003, 68, 9948–9957.
These authors added 2.0 equiv of a Grignard to the sulfinyl imine. We added
the sulfinyl imine to 4 equiv of Grignard . For other examples of sulfinyl
imine formation from an R-oxygenated aldehyde, see: (b) Barrow, J. C.; Ngo,
P. L.; Pellicore, J. M.; Selnick, H. G.; Nantermet, P. G. Tetrahedron Lett.
2001, 42, 2051–2054. (c) Tang, T. P.; Volkman, S. K.; Ellman, J. A. J. Org.
Chem. 2001, 66, 8772–8778. (d) Rech, J. C.; Floreancig, P. E. Org. Lett. 2003,
9, 1495–1498.
(20) Ahn, N. T. Top. Curr. Chem. 1980, 88, 145–162.
(21) Evans, D. A.; Siska, S. J.; Cee, V. J. Angew. Chem., Int. Ed. 2003, 42,
1761–1765.
(15) Kolb, H. C.; VanNieuwenhze, M. S.; Sharpless, K. B. Chem. Rev.
1994, 94, 2483–2547.
5976 J. Org. Chem. Vol. 74, No. 16, 2009

Harried et al.
JOCArticle
SCHEME 3. Gram-Scale Synthesis of Glyceraldehyde Synthon 11
SCHEME 4. Synthesis of Common Intermediate 6 and Stereoselective Addition of BnMgCl
98:2 ratio. Freshly prepared aldehyde 11 was subjected to
CuSO4-promoted22 imine formation with (S)-(-)-tert-butyl-
sulfinamide in DCM to give good yields (50-80%) of the di-
O-TBS protected N-tert-butanesulfinyl imine 6 (Scheme 4).
We found it necessary to purify the crude mixture by silica gel
chromatography in order to prevent decomposition of the
product. We did not observe any epimerization of the R-
silyloxy stereocenter during the formation of this imine. The
chiral auxiliary nicely provided a convenient handle for
simple 1D NMR analysis of the diastereomeric purity of 6
or 14. If the R-silyloxy stereocenter of the labile aldehyde 11
were to epimerize, then we could detect it by NMR. In fact,
CuSO4-promoted procedure was easier to perform than the
titanium-mediated methods.
Following Ellman’s optimized conditions,19 we subjected
6 to excess BnMgCl in THF with TMEDA at -78 °C to
obtain a 5.5:1.0 ratio of diastereomers in a 93% yield.24
Analysis of the crude reaction mixture in order to determine
the diastereomeric ratio of products in a proper fashion was
found to be very difficult. After much effort, we simply could
not discover an analytical reverse phase HPLC method that
would provide adequate resolution of the peaks. The pro-
ducts are very nonpolar and are retained on a reverse phase
column until 100% acetonitrile is used as the mobile phase,
resulting in broad peak shapes and poor resolution. The
major diastereomer 12a (higher Rf) was isolated in a 80%
yield after silica gel chromatography. The minor diastereo-
mer 13a (lower Rf) was easily separated from the major
diastereomer (ΔRf = 0.25). Subsequent experiments proved
that the major diastereomer 12a has the desired 2,3-anti
configuration for further elaboration to anti-N-BOC 3-ami-
no-1,2-epoxides 2a. To determine the effect of the chirality
on the sulfur atom toward the diastereofacial selectivity
of nucleophilic addition, the diastereomeric di-O-TBS
1
we synthesized the enantiomer of 14 (ent-14), obtained H
and 13C NMR spectra of a 1:1 mixture of ent-14 to 6, and
clearly observed doubling of carbon peaks. We also formed 6
using 2.5 equiv of Ti(OEt)4 in THF at room temperature for
2-14 h followed by addition to quickly stirred brine and
EtOAc with a filtration through Celite.23 In our hands, the
titanium-mediated process gave comparable yields
(50-70%) but resulted in somewhat difficult product isola-
tion due to formation of titanium dioxide emulsions. Ulti-
mately, we decided that the operational simplicity of the
(22) See refs 8 and 19.
(24) The diasteromeric ratio was determined by the mass balance of
isolated pure diasteromers. The ratio was confirmed by reverse phase HPLC
analysis (integration at 215 nm) of the resulting diols after treatment of the
crude reaction mixture to excess TBAF in THF for 24 h. Analysis of the crude
reaction mixture gave unsatisfactory results because the products are very
nonpolar and are not eluted until 100% acetonitrile is used as the mobile
phase, resulting in broad peak shapes and poor resolution.
(23) For representative example of Ti-mediated processes, see:
(a) Beenen, M. A.; An, C.; Ellman, J. A. J. Am. Chem. Soc. 2008, 130,
6910–6911. (b) Hjelmgaard, T.; Faure, S.; Lemoine, P.; Viossat, B.; Aitken,
D. J. Org. Lett. 2008, 10, 841–844. (c) Denolf, B.; Leemans, E.; De Kimpe, N.
J. Org. Chem. 2007, 72, 3211–3217. (d) Denolf, B.; Mangelinckx, S.; Trnroos,
K. W.; De Kimpe, N. Org. Lett. 2006, 8, 3129–3132. (e) Reference 19 .
J. Org. Chem. Vol. 74, No. 16, 2009 5977

Harried et al.
JOCArticle
SCHEME 5. Deprotection to Multifunctional Intermediate 18 and Synthesis of anti-N-BOC-3-Amino-1,2-epoxides 2a
protected N-tert-butanesulfinyl imine 14 was synthesized by
the condensation of aldehyde 11 with (R)-(+)-tert-butylsul-
finamide. The same conditions of excess BnMgCl in THF
with TMEDA at -78 °C were used on substrate 14 to give a
3.5:1.0 dr of 15 to 16 in a 67% yield. Interestingly, both
diastereomeric di-O-TBS protected N-tert-butanesulfinyl
imines 6 and 14 gave the 2,3-anti product as the major
product, albeit in slightly different ratios. Under the condi-
tions of excess BnMgCl in THF with TMEDA at -78 °C,25
the R-OTBS stereocenter is the major stereochemical deter-
minate toward the diastereofacial selectivity of nucleophilic
addition to the sulfinyl imines 6 and 14. The temperature,
solvent, concentration, solvent ratio of mixed solvents, and
chelating additives such as TMEDA all have an impact on
the diastereoselectivity of nucleophilic addition to sulfinyl
imines 6 and 14. It is important to follow the conditions of
THF (as the only or primary solvent), temperature, and
TMEDA closely, otherwise the diastereoselectivity may
erode or even favor the opposite diastereomer. Under Ell-
man’s conditions, the (S)-(-)-tert-butylsulfinamide is a
matched case that reinforces the Felkin-Ahn or Cornforth
models. The (R)-(+)-tert-butylsulfinamide resulted in a
mismatched diastereofacial bias.
Efficient functional group manipulations of the Felkin-
Ahn addition product 12a toward useful multifunctional
intermediates, including the desired anti-N-BOC 3-amino-
1,2-epoxides 2a, remained as our last synthetic challenge
before going into rapid production of P1 analogue epoxides.
The N-sulfinyl imine addition product 12a has three func-
tional groups that are potentially labile to acidic hydrolysis
conditions: the sulfinyl group, the primary TBS ether, and
the secondary TBS ether. Careful analyses of the reaction
rates were required to achieve the desired deprotections. The
sulfinyl group was the first functional group to be removed
when 12a was exposed to g4 equiv of HCl (4 N in 1,4-
dioxane) in EtOH at 0 °C (Scheme 5).
advantageous effect toward selective TBS ether ethanolysis.
Once the primary amine is protonated, the resulting positive
charge decreased the rate of adjacent secondary TBS depro-
tection. The qualitative rate of ethanolysis of the primary
TBS was found to be independent of both the equivalence of
HCl and the protonation state of the primary amine. We
were able to exploit these rate differences and observed
nearly complete chemoselective hydrolysis of both the sulfi-
nyl group and the primary TBS ether in the presence of the
secondary TBS ether to give the unisolated intermediate 17
after 5 h of HCl treatment at 0 °C.26 The amino alcohol 17 is a
useful and multifunctional molecule. The primary amine of
17 was conveniently protected as a BOC carbamate in 75-
85% yield by adding excess TEA followed by DCM and
BOC2O. The versatile intermediate 18a was converted to
anti-N-BOC 3-amino-1,2-epoxides 2a by activation of the
primary alcohol with MsCl and TEA in DCM. The crude
mesylate was immediately subjected to excess TBAF to give
the desired epoxide 2a in 80% yield for the two steps. The
experimental data for 2a matches that from the literature27
and therefore proves the configuration of the C.3 stereo-
center set during the addition of BnMgCl to imine 6.
Analytical supercritical fluid chromatography (SFC) using
a chiral column (ADH) along with commercially available 2a
and ent-2a (the 2R,3R-epoxide) as standards showed that
synthetic 2a was in fact the 2S,3S-epoxide. Moreover, the
enantiomeric excess of synthetic 2a is greater than 95%.
A paradigm of medicinal chemistry is to synthesize high
quality new chemical species in high purity in a short period
of time. One way to help accomplish this goal is to stockpile a
common intermediate and introduce diversity as late into the
synthesis of new analogues as possible. With the required
chemistry in place, we started to make analogues of anti-N-
BOC 3-amino-1,2-epoxides. Various organometallic re-
agents were added into the common intermediate 6 under
the same conditions of THF/TMEDA at -78 °C to give
good yields of single diastereomers 12a-i after silica gel
column chromatography (Table 1). All of the organometallic
reagents that were added to sulfinyl imine 6 gave the
Felkin-Ahn diastereomer 12 as the major product except
Under these reaction conditions, the sulfinyl group was
completely removed to give the primary amine between 15
and 60 min, according to analytical HPLC/MS analysis.
The deprotected primary amine was instantly protonated
in the acidic EtOH. This proton transfer event had an
(26) For a review on selective deprotection of silyl ethers, see: Nelson, T.
D.; Crouch, R. D. Synthesis 1996, 1031–1069.
(27) (a) Evans, B. E.; Rittle, K. E.; Homnick, C. F.; Springer, J. P.;
Hirshfield, J.; Veber, D. F. J. Org. Chem. 1985, 50, 4615–4625. (b) Beaulieu,
P. L.; Wernic, D. J. Org. Chem. 1996, 61, 3635–3645. (c) Ojima, I.; Wang, H.;
Wang, T.; Ng, E. W. Tetrahedron Lett. 1998, 39, 923–9126. (d) Reference 4.
(25) We found that minor changes to Ellman’s conditions gave rise to
different diastereomeric ratios. From a practical point of view, these changes
were not significant for our purposes. In every case, the desired diastereomer
was the higher Rf product. Moreover, the minor product was easily separated
by chromatography.
5978 J. Org. Chem. Vol. 74, No. 16, 2009

Harried et al.
JOCArticle
TABLE 1. Addition of Organometallics to 6 and Transformation to Epoxide 2a
aAll addition reactions were conducted in 4 equiv of RM in THF with 5 equiv of TMEDA at -78 °C. bRatios were determined by mass balance of
material except for entry 10, which was determined by analysis of crude 1H NMR. After many attempts, we could not create a satisfactory analytical
reverse-phase HPLC method to analyze the crude reaction mixtures for dr determination. cYield of 2a-i over three steps. dYield for diastereomer 20 not
2b. eReaction conducted at -45 °C.
for vinylmagnesium chloride. The syn-2,3 diastereomer 13b,
which is consistent with a chelation-controlled transition
state, was obtained in a slight excess when vinylmagnesium
chloride was added to the imine 6 under the standard
conditions of THF with TMEDA. The diasteromeric ratio
of 12 to 13 was excellent (>9:1) for the addition of phenyl
lithium, tert-butyl lithium, and isopropyl lithium (entries 3,
6, and 7).
For these three organolithium reagents, none of the minor
diastereomers 13c, 13f, and 13g were detected by either thin
layer chromatography (the products and starting imine
stained very well and in all cases the diastereomers 12 and
13 had a Rf difference of 0.25) or during silica gel chroma-
tography. The diastereoselectivity of MeLi addition to imine
6 was poor. The diastereoselectivity in favor of the desired
Felkin-Ahn product 12h was slightly enhanced by switching
from MeLi to MeMgCl. The rate of addition of MeMgCl to
imine 6 was much slower than the rate with MeLi, which
completely consumed the imine 6 in less than 1 h at -78 °C.
In fact, the addition of MeMgCl required elevated tempera-
ture (-45 °C) for 48 h and the imine was not completely
consumed. The remaining entries in Table 1 gave good
diastereomeric ratios of 12 to 13. In all cases, the desired
anti-2,3 addition product 12 was the higher Rf diastereomer.
Moreover, the minor diastereomer 13 (lower Rf) was easily
separated from the major diastereomer 12 (ΔRf = 0.25) by
silica gel chromatography. The optimized conditions for the
elaboration to the key coupling epoxide 2a were successfully
used for all of the sulfinyl imine addition products 12a-i.
This demonstrated the generality of our acidic ethanolysis
conditions. The R-chiral center is prone to racemization in
some of the literature routes toward anti-N-BOC 3-amino-
1,2-epoxides, such as 2a.28 Our method circumvents this
problem of unwanted racemization. The epoxide 2c29 has
appeared in the literature. Our method seems to be well
suited for making aryl P1 epoxides such as 2c.
Stereochemical proofs were conducted for all the epoxides
2 that were not previously known in the literature and for
epoxides 2g30 and 2h31 that were previously published com-
pounds (Scheme 6). This was accomplished by converting the
(28) See ref 4.
(29) To our knowledge, 2c has not been synthesized starting from
phenylglycine. However, 2c has been synthesized from other starting materi-
als: (a) Badorrey, R.; Cativiela, C.; Diaz-de-Villegas, M., D.; Galvez, J., A.
Tetrahedron 1999, 55, 14145–14160. (b) Reference 31b and 31c.
(30) (a) Rotella, D. P. Tetrahedron Lett. 1995, 36, 5453–5456.
(b) Reference 27b.
(31) (a) Reference 30a. (b) Castejon, P.; Moyano, A.; Pericas, M. A.; Riera, A.
Tetrahedron 1996, 52, 7063–7086. (c) Castejon, P.; Moyano, A.; Pericas, M. A.;
Riera, A. Tetrahedron Lett. 1995, 36, 3019–3022. (d) Reference 27b .
J. Org. Chem. Vol. 74, No. 16, 2009 5979

Harried et al.
JOCArticle
SCHEME 6. Conversion of Epoxides to Oxazolidinones; Stereo-
chemistry Determined by COSY and NOESY
Experimental Section
(R)-2,3-Dihydroxypropyl 4-Methoxybenzoate (8). Note: the
product of this reaction is thermally and hydrolytically sensitive.
Perform all workup procedures below 30 °C and removal of
solvents on a rotary evaporator below 40 °C. To a 2.0 L round-
bottom flask containing potassium hexacyanoferrate(III)
(68.988 g, 187.29 mmol) was added water (300 mL), and the
mixture was allowed to stir at 23 °C for 5 min. At this time,
potassium carbonate (25.885 g, 187.29 mmol) and (DHQ)2-
PHAL (0.48633 g, 0.62431 mmol) were added in one portion.
The mixture was allowed to stir for 15 min in order to allow all of
the salts to completely dissolve. At this point, tert-butanol (200
mL) was added, and the flask was placed in a 0 °C bath and
allowed to stir for 10 min before the addition of potassium
osmate dihydrate (0.23003 g, 0.62431 mmol) in one portion. The
allyl 4-methoxybenzoate (12.000 g, 62.431 mmol) was added in
tert-butanol (100 mL), and the internal temperature of the
reaction was determined to be 8 °C. The reaction was allowed
to stir for 2 h (the internal temp ranged from 3 to 8 °C during the
reaction) and then quenched by the addition of sodium thiosul-
fate (118.45 g, 749.18 mmol). The quenched reaction was
allowed to stir in the cooling bath for 10 min, then the bath
was removed, and the quenched reaction was allowed to stir for
10 more min before it was poured into EtOAc (300 mL) and
water (200 mL). The aq layer was extracted with EtOAc (3 Â 200
mL). The combined organics were washed with brine and dried
with sodium sulfate. The dried solution was passed through a
short plug of silica gel and concentrated to give 14.000 g of a
white solid in a 99% yield. Rf=0.30 in EtOAc; 1H NMR (400
MHz, CDCl3) δ ppm 3.60-3.80 (m, 2 H), 3.88 (s, 3 H), 4.00-
4.13 (m, 1 H), 4.35-4.55 (m, 2 H), 6.92 (d, J=8.80 Hz, 2 H), 8.00
(d, J=8.80 Hz, 2 H).
epoxides 2 to oxazolidinones32 19 where the cyclic structures
would allow for 2D NMR experiments (COSY and NOESY)
to clearly establish the relative stereochemistry of the hetero-
cyclic ring. The epoxides 2 were regioselectively opened at the
primary carbon by the action of thiophenol and sodium
hydride in DMF. The desired oxazolidinones were then
formed by heating the thiophenol addition adducts or remov-
ing the BOC group followed by treatment with triphosgene.
(S)-2,3-Bis(tert-butyldimethylsilyloxy)propyl 4-Methoxyben-
zoate (9). To a 1.0 L round-bottom flask containing (R)-2,3-
dihydroxypropyl 4-methoxybenzoate (4.800 g, 21.2 mmol) was
added DCM (100 mL), and the mixture was allowed to stir at
0 °C for 5 min. At this time, TEA (8.87 mL, 63.7 mmol) was
added, and the reaction was allowed to stir for 5 min before the
dropwise addition of tert-butyldimethylsilyl triflate (10.2 mL,
44.6 mmol) via a syringe. The reaction was allowed to stir for 1 h
and then quenched by pouring into HCl (0.1 N, 100 mL). The aq
layer was extracted with DCM (2 Â 75 mL). The combined
organics were washed with HCl (0.1 N, 2 Â 150 mL), sodium
bicarbonate (1 Â 150 mL, sat), and brine and dried with sodium
sulfate. The dried solution was passed through a plug of silica gel
to and concentrated to give 9.20 g of 9 as a colorless oil in 95%
yield. 1H NMR (300 MHz, CDCl3) δ ppm 0.04-0.14 (m, 12 H),
0.88 (s, 9 H), 0.90 (s, 9 H), 3.64 (d, J=5.87 Hz, 2 H), 3.87 (s, 3 H),
4.03 (dd, J=5.67, 4.50 Hz, 1 H), 4.25 (dd, J=6.0, 6.00 Hz, 1 H),
4.41 (dd, J=11.35, 3.91 Hz, 1 H), 6.93 (d, J=8.61 Hz, 2 H), 8.01
(d, J=8.41 Hz, 2 H); 13C NMR (75 MHz, CDCl3) δ -5.5, -5.4,
-4.8, -4.6, 18.0, 18.3, 25.7, 25.9, 55.3, 64.8, 66.3, 71.3, 113.5,
122.7, 131.6, 163.2, 166.2; HRMS (TOF) calcd for [C23H42O5Si2 +
H] 455.2644, found 455.2643, error=-0.13 ppm. Rf =0.75 in
20% EtOAc in hexanes, UV active and stains white to anisalde-
hyde stain. [R]34.0D=-15.0 (c=4.84 in CHCl3). FTIR (thin film)
Conclusions
In summary, a di-OTBS protected glyceraldehyde synthon
11 was synthesized using a Sharpless AD reaction in order
to establish the absolute stereochemistry in high opti-
cal purity. This aldehyde was condensed with Ellman’s
(S)-(-)-tert-butylsulfinamide reagent to form a bench-stable
sulfinyl imine 6 on a multigram scale. The sulfinyl imine 6
served as a common intermediate for the stereoselective
addition of various organometallic reagents in good yield.
The R-silyloxy stereocenter of 6 controlled the diastereofacial
selectivity of nucleophilic addition and is consistent with a
Felkin-Ahn or Cornforth model. The (S)-(-)-tert-butylsul-
finyl imine 6 provided a matched stereodifferentiation,
whereas the (R)-(+)-tert-butylsulfinyl imine 14 gave a mis-
matched result. The organometallic addition products were
efficiently converted to anti-N-BOC 3-amino-1,2-epoxides in
three steps. These anti-N-BOC 3-amino-1,2-epoxides are key
intermediates toward many commercial syntheses of HIV
protease inhibitors. Novel anti-N-BOC 3-amino-1,2-epoxides
could be used to synthesize a novel HIV protease inhibitor.
2955, 2930, 1719, 1607, 1512, 1257, 1168, 1101, 836, 777 cm-1
.
(S)-2,3-Bis(tert-butyldimethylsilyloxy)propan-1-ol (10). To a
3.0 L round-bottom flask containing (S)-2,3-bis(tert-butyldim-
ethylsilyloxy)propyl 4-methoxybenzoate (26.5000 g, 58.3 mmol)
was added DCM (300 mL), and the mixture was allowed to stir
at -78 °C for 15 min. At this time, DIBAL-H (1.0 M in hexanes)
(117 mL, 117 mmol) was added via syringe over 20 min. The
reaction was allowed to stir an additional 30 min, when TLC
indicated that all starting material was consumed, and then
quenched by the addition of methanol (7.09 mL, 175 mmol) via
syringe. The quenched reaction was allowed to stir for 5 min, and
(32) (a) Luly, J. R.; Dellaria, J. F.; Plattner, J. J.; Soderquist, J. L.; Yi, N.
J. Org. Chem. 1987, 52, 1487–1492. (b) Barrett, A. G. M.; Seefeld, M. A.;
White, A., J. P.; Williams, D. J. J. Org. Chem. 1996, 61, 2677–2685.
5980 J. Org. Chem. Vol. 74, No. 16, 2009

Harried et al.
JOCArticle
then potassium sodium tartrate (411 g, 1457 mmol) in 250 mL of
water was added. The reaction was allowed to warm to 23 °C and
stirred for 2 h when the layers nicely separated. The aq layer was
extracted with DCM (3 Â 150 mL). The combined organics were
washed with brine and dried with sodium sulfate over the weekend.
The dried solution was filtered and concentrated to give 28 g of
crude oil. This oil was passed through a plug of silica gel and eluted
with 30% EtOAc in hexanes (four 1.0 L fractions, product in
fractions 2 and 3) to remove the 4-methoxy benzyl alcohol
byproduct (Rf = 0.10 in 20% EtOAc in hexanes, stains red to
anisaldehyde, and is UV active) and give 18.40 g of primary
alcohol 10 in 98% yield. Product Rf = 0.50 in 20% EtOAc in
hexanes, not UV active, faint gray to anisaldehyde stain. 1H NMR
(300 MHz, CDCl3) δ ppm 0.03-0.20 (m, 12 H) 0.81-0.95 (m,
18 H) 3.45-3.84 (m, 5 H); 13C NMR (75 MHz, CDCl3) δ -5.51,
-5.46, -4.8, -4.7, 18.0, 18.3, 25.8, 25.9, 64.8, 64.9, 72.5; HRMS
(TOF) calcd for [C15H36O3Si2 + H] 321.2276, found 321.2276,
error=0.06 ppm. Rf=0.50 in 20% EtOAc in hexanes, UV active
and faint gray to anisaldehyde. [R]34.5D=-13.0 (c=2.2 in CHCl3).
FTIR (thin film) 3445, 2956, 2929, 2858, 1473, 1256, 1093, 836,
25.65, 25.80, 56.59, 66.03, 75.32, 169.52; HRMS (TOF) calcd for
[C19H43NO3Si2 + H] 422.2575, found 422.2581, error =1.45
ppm. Rf=0.50 in 20% EtOAc in hexanes, UV active and stains
yellow to anisaldehyde. [R]34.4D =+166.0 (c=3.2 in CHCl3).
FTIR (thin film) 2956, 2930, 2859, 1625, 1473, 1254, 1142, 1091,
835, 779 cm-1
.
(R,E)-N-((S)-2,3-Bis(tert-butyldimethylsilyloxy)propylidene)-
2-methylpropane-2-sulfinamide (14). The same procedure used
to synthesize compound 6 was followed, and 2.1 g of a colorless
oil was isolated (50% yield). 1H NMR (300 MHz, CDCl3) δ ppm
0.03-0.11 (m, 12 H), 0.88 (s, 9 H), 0.90 (s, 9 H), 1.21 (s, 9 H),
3.74-3.78 (m, 2 H), 4.44-4.53 (m, 1 H), 7.99 (d, J=3.95 Hz,
1 H); 13C NMR (75 MHz, CDCl3) δ -5.43, -5.41, -4.83, -
4.71, 18.19, 18.31, 22.39, 25.71, 25.78, 56.78, 65.93, 75.44,
169.48; HRMS (TOF) calcd for [C19H43NO3Si2
+ H]
422.2575 found 422.2589, error=3.25 ppm. Rf=0.30 in 10%
EtOAc in hexanes, UV active and stains pink to anisaldehyde.
[R]34.4D=-112.0 (c=2.35 in CHCl3). FTIR (thin film) 2956,
2930, 2859, 1627, 1473, 1256, 1091, 836, 778 cm-1
.
General Procedure for the Grignard or Organolithium Addition
to N-tert-Butanesulfinyl Imine 6 or 14. To a 1.0 L round-bottom
flask containing the Grignard reagent (in THF) (24.9 mL, 49.8
mmol) was added THF (100 mL), and the mixture was allowed
to stir at -78 °C for 15 min. At this time, TMEDA (9.39 mL,
62.2 mmol) was added via syringe, and the reaction was allowed
to stir for 30 min prior to the addition of sulfinyl imine (5.25 g,
12.4 mmol) in THF (15 mL) dropwise over 15 min. The reaction
was allowed to stir for 5 h and then quenched by the addition of
ammonium chloride (satd 150 mL). The aq layer was extracted
with EtOAc (3 Â 100 mL). The combined organics were washed
with brine, dried with sodium sulfate, filtered, and concentrated
to give 7 g of a crude oil. The crude oil was subjected to a 330 g
Isco column (10 to 35%EtOAc in hexanes) to give pure single
diastereomeric product.
777 cm-1
.
(R)-2,3-Bis(tert-butyldimethylsilyloxy)propanal (11). To a 500
mL round-bottom flask containing (S)-2,3-bis(tert-butyldim-
ethylsilyloxy)propan-1-ol 10 (4.70 g, 14.7 mmol) was added
DCM (100 mL), and the mixture was allowed to stir at 23 °C
for 2 min. At this time, sodium bicarbonate (3.69 g, 44.0 mmol)
and Dess-Martin periodinane (7.46 g, 17.6 mmol) were added
in one portion, and the reaction was allowed to stir for 1.5 h. The
reaction was quenched by the addition of sodium thiosulfate
(6.95 g, 44.0 mmol) in one portion followed by sodium bicarbo-
nate (satd 250 mL) and diethyl ether (250 mL). The quenched
reaction was allowed to stir for 45 min and then the clear layers
were separated. The organic layer was washed with sodium
bicarbonate (2 Â 250 mL), water (1 Â 250 mL) and brine (1 Â
100 mL). The organic layer was dried with magnesium sulfate,
filtered, and concentrated to give 5.00 g of a colorless oil. Rf=
0.80 in 20% EtOAc in hexanes, not UV active, stains pink/
orange to anisaldehyde. The aq layers were back extracted with
ether (2 Â 125 mL). The combined back extractions were washed
with brine, dried with magnesium sulfate, filtered, and concen-
trated to give less than 200 mg of oil. 1H NMR (300 MHz,
CDCl3) δ ppm 0.06 (s, 3 H), 0.06 (s, 3 H), 0.09 (s, 3 H), 0.10 (s, 3
H), 0.88 (s, 9 H), 0.92 (s, 9 H), 3.81 (d, J=5.26 Hz, 2 H), 4.06 (dt,
J=5.33, 1.32 Hz, 1 H), 9.66 (d, J=1.32 Hz, 1 H); 13C NMR (75
MHz, CDCl3) δ -5.50, -5.47, -4.84, -4.80, 18.26, 18.28, 25.74,
25.81, 64.64, 78.85, 203.22; HRMS (TOF) calcd for
[C15H34O3Si2 + H] 319.2119, found 319.2134, error = 4.54
ppm. Rf = 0.80 in 20% EtOAc in hexanes, not UV, stains
(S)-N-((2S,3S)-3,4-Bis(tert-butyldimethylsilyloxy)-1-phenyl-
butan-2-yl)-2-methylpropane-2-sulfinamide (12a). Higher Rf spot
is the major diastereomer. The material was subjected to high
vacuum overnight to give 5.07 g of clear oil. 1H NMR (300 MHz,
CDCl3) δ 0.08-0.12 (m, 9H), 0.17 (s, 3H), 0.93 -0.55 (m, 27H),
2.59-2.67 (m, 1H), 2.88-2.94 (m, 1H), 3.53-3.59 (m, 2H),
3.63-3.69 (m, 1H), 3.74-3.82 (m, 1H), 4.13-4.19 (m, 1H),
7.13-7.17 (m, 3H), 7.22-7.27 (m, 2H); 13C NMR (75 MHz,
CDCl3) δ -5.5, -5.4, -4.5, -4.2, 18.1, 18.3, 22.3, 25.9, 25.9,
35.1, 55.7, 60.7, 64.3, 75.6, 125.9, 128.1, 129.5, 139.4; HRMS
(TOF) calcd for [C26H51NO3SSi2 + H] 514.3201, found
514.3187, error = -2.68 ppm. Rf = 0.40 in 20% EtOAc in
hexanes, UV active and stains purple to anisaldehyde. [R]34.4
=
D
-24.0 (c=3.16 in CHCl3). FTIR (thin film) 2954, 2930, 2858, 1472,
pink/orange to anisaldehyde. [R]31.5 = +6.0 (c = 2.67 in
1254, 1127, 1076, 837, 779, 698 cm-1
.
D
CHCl3). FTIR (thin film) 2957, 2930, 2859, 1740, 1473, 1254,
.
General Procedure for Deprotection of 12 to Primary Alcohols
18: tert-Butyl (2S,3S)-3-(tert-butyldimethylsilyloxy)-4-hydroxy-
1-phenylbutan-2-ylcarbamate (18a). To a 1.0 L round-bottom
flask containing 12 (9.73 mmol) was added EtOH (50 mL), and
the mixture was allowed to stir at 0 °C for 10 min. At this time,
HCl (4.0 M in 1,4-dioxane) (9.73 mL, 38.9 mmol) was added in
one portion. After 5 h LC/MS showed complete removal of
sulfinyl and primary silyl ether, TEA (6.78 mL, 48.6 mmol) and
DCM (20 mL) were added, and the reaction was removed from
the ice bath. After 10 min, BOC2O (3.19 g, 14.6 mmol) was
added in one portion, and the reaction was allowed to stir for
14 h. The bulk of the solvent was removed, and then the residue
was dissolved in EtOAc (300 mL). The organic layer was washed
with ammonium chloride (3 Â 100 mL), sodium bicarbonate,
and brine. The organic layer was dried with sodium sulfate,
filtered, and concentrated to give 18a as an oil. The crude
product was subjected to a 330 g Isco column (20% to 45%
EtOAc in hexanes) to give 3.40 g of colorless oil (88% yield). 1H
NMR (300 MHz, CDCl3) δ ppm -0.00 (s, 3 H), 0.01 (s, 3 H),
1115, 978, 836, 779 cm-1
(S,E)-N-((S)-2,3-Bis(tert-butyldimethylsilyloxy)propylidene)-
2-methylpropane-2-sulfinamide (6). To a 500 mL round-bottom
flask containing (R)-2,3-bis(tert-butyldimethylsilyloxy)pro-
panal 11(4.670 g, 14.7 mmol) was added DCM (100 mL), and
the mixture was allowed to stir at 23 °C for 2 min. At this time,
(S)-2-methylpropane-2-sulfinamide (2.13 g, 17.6 mmol) and
copper(II) sulfate (5.85 g, 36.6 mmol) were added, and the
reaction was allowed to stir for 24 h. At this time TLC showed
that all of the aldehyde was consumed. The crude reaction
mixture was filtered through a plug of Celite in order to remove
the solid copper salt. The organic layer was concentrated to give
7.50 g of oil, which was subjected to a 330 g Isco column (10-
35% EtOAc in hexanes) to give 4.90 g of a white solid. 1H NMR
(300 MHz, CDCl3) δ ppm 0.05 (s, 6 H), 0.08 (s, 3 H), 0.09 (s, 3
H), 0.88 (s, 9 H), 0.90 (s, 9 H), 1.20 (s, 9 H), 3.67-3.81 (m, 2 H),
4.43-4.53 (m, 1 H), 7.98 (d, J=4.33 Hz, 1 H); 13C NMR (75
MHz, CDCl3) δ -5.54, -5.49, -4.84, -4.78, 18.09, 18.28, 22.27,
J. Org. Chem. Vol. 74, No. 16, 2009 5981

Harried et al.
JOCArticle
0.03 (s, 1 H), 0.84 (s, 9 H), 1.25 (s, 9 H), 2.44-2.73 (m, 2 H), 2.91
(dd, J=13.74, 4.38 Hz, 1 H), 3.42-3.70 (m, 3 H), 3.79-4.03 (m,
1 H), 4.63 (d, J=8.77 Hz, 1 H), 6.96-7.32 (m, 5 H)); 13C NMR
(75 MHz, CDCl3) δ -4.8, -4.4, 18.1, 25.4, 28.4, 36.5, 54.0, 63.6,
73.9, 79.4, 126.7, 128.4, 129.1, 138.3, 155.9.; HRMS (TOF) calcd
for [C21H37NO4Si + H] 396.2565, found 396.2573, error=1.98
ppm. Rf=0.25 in 20% EtOAc in hexanes, UV active and stains
white to anisaldehyde stain.
and concentrated to give a white solid that was subjected to a
40 g Isco column (15 to 45% EtOAc in hexanes) to give 130 mg
of a white solid (80% yield). 1H NMR (300 MHz, CDCl3) δ 1.38
(s, 9H), 2.73-3.00 (m, 5H), 3.69 (br s, 1H), 4.49 (br s, 1H), 7.21-
7.34 (m, 5H); 13C NMR (75 MHz, CDCl3) δ 28.2, 46.8, 52.5,
53.2, 79.6, 126.6, 128.5 129.4, 136.7, 155.2; HRMS (ESI) calcd
for [C15H21NO3 + Na] 286.1414, found 286.1420. Rf=0.40 in
30% EtOAc in hexanes, UV active and stains yellow/brown to
anisaldehyde stain, cospots with commercially available materi-
al. [R]34.3D=+6.0 (c=1.0 in CHCl3). FTIR (KBr pellet) 3378,
2982, 1680, 1524, 1170, 928, 702 cm-1. Mp=121-123 °C, lit=
122-124 °C.27b
Representative Procedure for Conversion of Primary Alcohols
18 to Epoxides 2: tert-Butyl (S)-1-((S)-Oxiran-2-yl)-2-phenyl-
ethylcarbamate (2a). To a 250 mL round-bottom flask containing
tert-butyl (2S,3S)-3-(tert-butyldimethylsilyloxy)-4-hydroxy-1-ph-
enylbutan-2-ylcarbamate 18a (310.00 mg, 784 μmol) was added
DCM (5 mL), and the mixture was allowed to stir at 0 °C for 10 min.
At this time, DIPEA (411 μL, 2351 μmol) and Ms-Cl (0.5 M
in DCM) (1567 μL, 784 μmol) were added via syringe. The
reaction was allowed to stir for 1 h and then poured into
ammonium chloride (satd 75 mL) and extracted with DCM
(4 Â 35 mL). The combined organics were washed with brine,
dried with sodium sulfate, passed though a short plug of silica gel,
and concentrated to give 370.00 mg of colorless oil. Rf=0.20 in
20% EtOAc in hexanes (slightly lower than starting primary
alcohol) UV active and stains white/pink to anisaldehyde).
The material was taken directly into the epoxide forming reaction.
To a 150 mL round-bottom flask containing the mesylate
(290.00 mg, 612.20 μmol) was added THF (10 mL), and the
mixture was allowed to stir at 0 °C for 10 min. At this time,
TBAF (192.08 mg, 734.64 μmol) was added in one portion, and
the reaction was allowed to stir for 14 h before it was transferred
to a separatory funnel with EtOAc (100 mL). The organic layer
was washed with water (100 mL). The aq layer was back
extracted with EtOAc (50 mL). The dried solution was filtered
Acknowledgment. We gratefully acknowledge and thank
Drs. Jason Tedrow, Filisity Vounatsos, and Kenneth McRae
and Mr. Bobby Riahi for helpful discussions, technical
assistance, and large scale syntheses. The authors thank
Drs. Randy Jensen, Steven Sukits, and Christopher Wilde
for conducting 2D NMR experiments, Steven Notari for the
development of analytical HPLC methods, Zheng Hua for
chiral SFC, and Dr. Paul Schnier for high resolution mass
spectra. We thank Dr. Ted Judd and Frenel DeMorin for
helpful discussions.
Supporting Information Available: General experimental
information and complete synthetic details and characterization
for compounds 7, 13a, 15, 16, 12b, 13b, 12a-i, 18a, 20, 2c-i, 19i,
19d, 19e, 19f, and 21; copies of 1H NMR and 13C NMR; copies
of 2D NMR experiments (COSY and NOESY) for compound
19d. This material is available free of charge via the Internet at
http://pubs.acs.org.
5982 J. Org. Chem. Vol. 74, No. 16, 2009