Asymmetric synthesis of chiral organofluorine compounds: Use of nonracemic fluoroiodoacetic acid as a practical electrophile and its application to the synthesis of monofluoro hydroxyethylene dipeptide isosteres within a novel series of HIV protease inhibitors

Article abstract of DOI:10.1021/ja010113y

Two stereoselective routes to a series of diastereomeric inhibitors of HIV protease, monofluorinated analogues of the Merck HIV protease inhibitor indinavir, are described. The two routes feature stereoselective construction of the fluorinated core subunits by asymmetric alkylation reactions. The first-generation syntheses were based on the conjugate addition of the lithium enolate derived-from pseudoephedrine α-fluoroacetamide to nitroalkene 12, a modestly diastereoselective transformation. A more practical second-generation synthetic route was developed that is based on a novel method for the asymmetric synthesis of organofluorine compounds, by enolate alkylation using optically active fluoroiodoacetic acid as the electrophile in combination with a chiral amide enolate. Resolution of fluoroiodoacetic acid with ephedrine provides either enantiomeric form of the electrophile in ≥96% ee. Alkylation reactions with this stable and storable chiral fluorinated precursor are shown to proceed in a highly stereospecific manner. With the development of substrate-controlled syn- or anti-selective reductions of α-fluoro ketones 44 and 45 (diastereomeric ratios 12:1-84:1), efficient and stereoselective routes to each of the four targeted inhibitors were achieved. The optimized synthetic route to the most potent inhibitor (syn,syn-4, K_i = 2.0 nM) proceeded in seven steps (87% average yield per step) from aminoindanol hydrocinnamide 40 and (S)-fluoroiodoacetic acid, and allowed for the preparation of more than 1 g of this compound. The inhibition of HIV-1 protease by each of the fluorinated inhibitors was evaluated in vitro, and the variation of potency as a function of inhibitor stereochemistry is discussed.

Full text of DOI:10.1021/ja010113y

J. Am. Chem. Soc. 2001, 123, 7207-7219
7207
Asymmetric Synthesis of Chiral Organofluorine Compounds: Use of
Nonracemic Fluoroiodoacetic Acid as a Practical Electrophile and Its
Application to the Synthesis of Monofluoro Hydroxyethylene
Dipeptide Isosteres within a Novel Series of HIV Protease Inhibitors
Andrew G. Myers,* Joseph K. Barbay, and Boyu Zhong
Contribution from the Department of Chemistry and Chemical Biology, HarVard UniVersity,
12 Oxford Street, Cambridge, Massachusetts 02138
ReceiVed January 12, 2001. ReVised Manuscript ReceiVed March 21, 2001
Abstract: Two stereoselective routes to a series of diastereomeric inhibitors of HIV protease, monofluorinated
analogues of the Merck HIV protease inhibitor indinavir, are described. The two routes feature stereoselective
construction of the fluorinated core subunits by asymmetric alkylation reactions. The first-generation syntheses
were based on the conjugate addition of the lithium enolate derived from pseudoephedrine R-fluoroacetamide
to nitroalkene 12, a modestly diastereoselective transformation. A more practical second-generation synthetic
route was developed that is based on a novel method for the asymmetric synthesis of organofluorine compounds,
by enolate alkylation using optically active fluoroiodoacetic acid as the electrophile in combination with a
chiral amide enolate. Resolution of fluoroiodoacetic acid with ephedrine provides either enantiomeric form of
the electrophile in g96% ee. Alkylation reactions with this stable and storable chiral fluorinated precursor are
shown to proceed in a highly stereospecific manner. With the development of substrate-controlled syn- or
anti-selective reductions of R-fluoro ketones 44 and 45 (diastereomeric ratios 12:1-84:1), efficient and
stereoselective routes to each of the four targeted inhibitors were achieved. The optimized synthetic route to
the most potent inhibitor (syn,syn-4, Ki ) 2.0 nM) proceeded in seven steps (87% average yield per step)
from aminoindanol hydrocinnamide 40 and (S)-fluoroiodoacetic acid, and allowed for the preparation of more
than 1 g of this compound. The inhibition of HIV-1 protease by each of the fluorinated inhibitors was evaluated
in vitro, and the variation of potency as a function of inhibitor stereochemistry is discussed.
Introduction
The identification of potassium fluoroacetate as the toxic
principle of the South African plant Dichapetalum cymosum in
1943 by Marais is often regarded as an important early discovery
that directed attention to the potential of fluorine substitution
to profoundly influence the biological activity of organic
molecules.1 Shortly thereafter, Fried reported the synthesis of
9R-fluorohydrocortisone acetate (1) and showed that it exhibited
glucocorticoid activity 11 times greater than that of the
corresponding hydrocarbon (cortisol acetate, 2).2 The enhanced
activity has been attributed to the lower pKa of the 11â-hydroxyl
group and its increased ability to donate a hydrogen bond, as
well as its increased resistance to metabolic inactivation by
oxidation. Fried’s discovery led to the development of at least
six commercial antiinflammatory agents. More importantly, his
demonstration of the extraordinary potential of fluorine substitu-
tion to alter and enhance the pharmacological properties of
organic molecules became the basis of a powerful strategy for
lead development in the pharmaceutical industry.
gen), however, fluorine substitution can alter the metabolic
stability, hydrogen-bonding capacity, lipophilicity, solubility,
conformation, and even the fundamental structure of a molecule
(e.g., the propensity of fluorinated ketones to hydrate), variations
that can profoundly influence biological activity.3 The impor-
tance of fluorine substitution in pharmaceutical development is
evident in the large number of fluorinated compounds that have
been approved by the FDA as drugs; fluorinated anticancer,
antiviral, antibacterial, antidiabetic, antimalarial, antifungal,
antidepressant, antipyschotic, antiinflammatory, and anesthetic
agents, among others, are represented.4
Because of the small size of the fluorine atom (van der Waals
radius 1.47 Å versus 1.20 Å for hydrogen), replacement of a
carbon-hydrogen bond with a carbon-fluorine bond does not
typically present a major steric perturbation in a molecule. As
a consequence of the unique physicochemical properties of
fluorine (Pauling electronegativity 3.98 versus 2.20 for hydro-
Analysis of an electronic database of patented pharmaceutical
substances reveals a surprising lack of structural diversity among
(3) For reviews on the physical properties of fluorinated compounds,
see: (a) Smart, B. E. In Chemistry of Organic Fluorine Compounds II: A
Critical ReView; Hudlicky, M., Pavlath, A. E., Eds.; ACS Monograph 187;
American Chemical Society: Washington, DC, 1995; pp 979-1010. (b)
Smart, B. E. In Organofluorine Chemistry: Principles and Commercial
Applications; Banks, R. E., Smart, B. E., Tatlow, J. C., Eds.; Plenum
Publishing Corporation: New York, 1994; pp 57-88.
(1) Marais, J. S. C. Onderstepoort J. Vet. Sci. Anim. Ind. 1943, 18, 203.
(2) Fried, J.; Sabo, E. F. J. Am. Chem. Soc. 1954, 76, 1455-1456.
10.1021/ja010113y CCC: $20.00 © 2001 American Chemical Society
Published on Web 07/07/2001

7208 J. Am. Chem. Soc., Vol. 123, No. 30, 2001
Myers et al.
the fluorinated pharmaceutical agents that have been developed
in the 47 years since Fried’s synthesis of 9R-fluorohydrocor-
tisone acetate (1).5 Among marketed fluorine-containing drugs
(107 compounds), only eight contain a stereogenic (sp3-
hybridized) carbon atom bearing fluorine. Of these, six are 9R-
fluorinated steroids, and the remaining two compounds, des-
flurane and enflurane, are fluorinated inhalation anesthetics
marketed as racemic mixtures. It is likely that a major factor
contributing to this lack of structural diversity is that there are
few practical methods for the asymmetric synthesis of chiral
organofluorine compounds. In particular, there are few methods
to form carbon-carbon bonds to fluorinated, stereogenic carbon
centers.6
Organofluorine compounds have been of particular impor-
tance in the development of inhibitors of carboxylic ester and
peptide hydrolases.7 Brodbeck first described the use of
fluorinated aldehydes and ketones as inhibitors of hydrolases,
demonstrating inhibition of acetylcholinesterase.8 Subsequently,
Abeles and co-workers showed that fluoroketones inhibited
proteases operating by widely varying mechanisms; aspartyl
proteases, zinc metalloproteases, and serine proteases were all
inhibited.9 They demonstrated that the fluoroketone inhibitors
form stable tetrahedral intermediates by the addition of active-
site nucleophiles to the electrophilic carbonyl and proposed that
these adducts mimic the transition state for peptide hydrolysis.10
Figure 1. Examples of fluorinated peptidomimetic protease inhibitors.
Fluorinated methyl ketones, inhibitors containing the difluo-
roketomethylene dipeptide isostere, and difluorostatone and
difluorostatine analogues are the most commonly studied classes
of fluorine-containing protease inhibitors today (Figure 1).
Protease inhibitors containing a fluorine-substituted stereogenic
carbon atom have been less frequently studied.11-14
In an extension of our recent efforts to develop new
methodology for the asymmetric synthesis of organofluorine
compounds using fluorocarbon-carbon bond-forming reac-
tions,15,16 we sought to apply our methodology in the synthesis
of a new class of structurally complex, fluorinated protease
inhibitors, specifically, inhibitors of HIV protease. HIV protease
is perhaps the most significant protease target in antiinfective
therapy today. Its inhibition has been shown to extend the length
and improve the quality of life of AIDS patients. The develop-
ment and large-scale production of the stereochemically complex
HIV protease inhibitor indinavir (Crixivan) by a team of Merck
scientists is one of the great recent achievements in the
pharmaceutical industry.17 Fluorine substitution of the hydroxy-
ethylene dipeptide isostere of indinavir is a logical consideration
in the search for inhibitors with improved potencies and
properties, but the indinavir structural platform illustrates well
(4) (a) Edwards, P. N. In Organofluorine Chemistry: Principles and
Commercial Applications; Banks, R. E., Smart, B. E., Tatlow, J. C., Eds.;
Plenum Publishing Corporation: New York, 1994; pp 501-541. (b) Ojima,
I., McCarthy, J. R., Welch, J. T., Eds. Biomedical Frontiers of Fluorine
Chemistry; ACS Symposium Series 639; American Chemical Society:
Washington, 1996. (c) Welch, J. T.; Eswarakrishnan, S. Fluorine in
Bioorganic Chemistry; John Wiley and Sons: New York, 1991. (d) Filler,
R. In Asymmetric Fluoroorganic Chemistry: Synthesis, Applications, and
Future Directions; Ramachandran, P. V., Ed.; ACS Symposium Series 746;
American Chemical Society: Washington, DC, 2000; pp 1-20. (e) For a
review focusing on marketed pharmaceuticals, see: Elliott, A. J. In
Chemistry of Organic Fluorine Compounds II: A Critical ReView; Hudlicky,
M., Pavlath, A. E., Eds.; ACS Monograph 187; American Chemical
Society: Washington, DC, 1995; pp 1119-1125. (f) O’Hagan, D.; Rzepa,
H. S. J. Chem. Soc., Chem. Commun. 1997, 645-652.
(5) The ISIS/Base MDL Drug Data Report (MDDR-3D) database was
searched for compounds containing a carbon-fluorine bond whose devel-
opmental phase was “Launched”. ISIS/Base, Version 2.2; Molecular Design
Ltd.; San Leandro, CA.
(6) For reviews and monographs on the asymmetric synthesis of chiral
organofluorine compounds, see (a) Soloshonok, V. A., Ed. Enantiocontrolled
Synthesis of Fluoro-Organic Compounds: Stereochemical Challenges and
Biomedicinal Targets; John Wiley and Sons Ltd.: New York, 1999. (b)
Ramachandran, P. V., Ed. Asymmetric Fluoroorganic Chemistry: Synthesis,
Applications, and Future Directions; ACS Symposium Series 746; American
Chemical Society: Washington, DC, 2000. (c) Kitazume, T.; Yamazaki,
T. Experimental Methods in Organic Fluorine Chemistry; Gordon and
Breach Science Publishers: Amsterdam, 1998. (d) Percy, J. M. Contemp.
Org. Synth. 1995, 2, 251-268. (e) Iseki, K.; Kobayashi, Y. In ReViews on
Heteroatom Chemistry; Oae, S., Ohno, A., Okuyama, T., Eds.; MYU:
Tokyo, 1995; Vol. 12, pp 211-237. (f) Welch, J. T. Tetrahedron 1987,
43, 3123-3197. (g) Percy, J. M. In Organofluorine Chemistry: Techniques
and Synthons; Chambers, R. D., Ed.; Topics in Current Chemistry, Volume
193; Springer-Verlag: Berlin, 1997; pp 131-195. (h) Mann, J. Chem. Soc.
ReV. 1987, 16, 381-436. (i) Hayashi, T.; Soloshonok, V, A. (Guest Editors).
Enantiocontrolled Synthesis of Fluoro-organic Compounds. Tetrahedron:
Asymmetry 1994, 5, 955-1126. (j) Resnati, G. Tetrahedron 1993, 49, 9385-
9445. (k) Bravo, P.; Resnati, G. Tetrahedron: Asymmetry 1990, 1, 661-
692.
(10) (a) Liang, T.-Y.; Abeles, R. H. Biochemistry 1987, 26, 7603-7608.
(b) Brady, K.; Wei, A.; Ringe, D.; Abeles, R. H. Biochemistry 1990, 29,
7600-7607.
(11) Garrett, G. S.; Emge, T. J.; Lee, S. C.; Fischer, E. M.; Dyehouse,
K.; McIver, J. M. J. Org. Chem. 1991, 56, 4823-4826.
(12) (a) Hoffman, R. V.; Saenz, J. E. Tetrahedron Lett. 1997, 38, 8469-
8472. (b) Hoffman, R. V.; Tao, J. Tetrahedron Lett. 1998, 39, 4195-4198.
(c) Hoffman, R. V.; Tao, J. J. Org. Chem. 1999, 64, 126-132. (d) Hoffman,
R. V.; Tao, J. In Asymmetric Fluoroorganic Chemistry: Synthesis,
Applications, and Future Directions; Ramachandran, P. V., Ed. ACS
Symposium Series 746; American Chemical Society: Washington, DC,
2000; pp 52-59.
(13) (a) Welch, J. T.; Abudellah, A. Tetrahedron: Asymmetry 1994, 5,
1005-1013. (b) Welch, J. T.; Araki, K.; Kawecki, R.; Wichtowski, J. A. J.
Org. Chem. 1993, 58, 2454-2462.
(14) (a) De Lucca, G. V.; Liang, J.; De Lucca, I. J. Med. Chem. 1999,
42, 135-152. (b) De Lucca, G. V.; Qi, H.; Jadhav, P. K.; Kassir, J. M.;
Lam, P. Y.-S.; McHugh, R. J., Jr. Substituted Cyclic Ureas and Derivatives
Thereof Useful as Retroviral Protease Inhibitors. Int. Pat. App. 97/08150,
March 6, 1997.
(7) A general review on protease inhibitors: (a) Rich, D. H. In
ComprehensiVe Medicinal Chemistry; Sammes, P. G., Ed.; Pergamon
Press: Oxford, U. K., 1990; pp 391-441. (b) For a specific review on
fluorinated inhibitors of aspartyl proteases, see: Sham, H. L. In Biomedical
Frontiers of Fluorine Chemistry; Ojima, I., McCarthy, J. R., Welch, J. T.,
Eds.; American Chemical Society: Washington, DC, 1996; pp 184-195.
(8) Brodbeck, U.; Schweikert, K.; Gentinetta, R.; Rottenberg, M. Biochim.
Biophys. Acta 1979, 567, 357-369.
(15) (a) Myers, A. G.; McKinstry, L.; Barbay, J. K.; Gleason, J. L.
Tetrahedron Lett. 1998, 39, 1335-1338. (b) Myers, A. G.; McKinstry, L.;
Gleason, J. L. Tetrahedron Lett. 1997, 38, 7037-7040. For the use of
pseudoephedrine as a chiral auxiliary in asymmetric alkylation reactions,
see: (c) Myers, A. G.; Yang, B. H.; Chen, H.; Gleason, J. L. J. Am. Chem.
Soc. 1994, 116, 9361-9362. (d) Myers, A. G.; Yang, B. H.; Chen, H.;
McKinstry, L.; Kopecky, D. J.; Gleason, J. L. J. Am. Chem. Soc. 1997,
119, 6496-6511.
(9) (a) Gelb, M. H.; Svaren, J. P.; Abeles, R. H. Biochemistry 1985, 24,
1813-1817. (b) Imperiali, B.; Abeles, R. H. Biochemistry 1986, 25, 3760-
3767.
(16) For other asymmetric fluorocarbon-carbon bond-forming reactions,
see: (a) Less, S. L.; Handa, S.; Millburn, K.; Leadlay, P. F.; Dutton, C. J.;
Staunton, J. Tetrahedron Lett. 1996, 37, 3515-3518. (b) Reference 13.

Synthesis of HIV Protease Inhibitors
J. Am. Chem. Soc., Vol. 123, No. 30, 2001 7209
Chart 1
the difficulty inherent in the stereospecific introduction of single
fluorine atoms within a complex target molecule: four diaster-
eomeric monofluoro hydroxyethylene dipeptide isosteres are
possible, and each represents a formidable and distinct synthetic
problem (structures 4, Chart 1).18 We recognized the challenge
that the synthesis of these targets represented and elected to
develop asymmetric routes to each of the four monofluorinated
indinavir analogues with the expectation that any viable solutions
that emerged would hold promise for application in a more
general sense.19
protease inhibitors. These integrate two new methods for the
asymmetric synthesis of fluorocarbon-carbon bonds, reported
here for the first time. This methodology was used to prepare
each of the diastereomeric indinavir analogues 4 in optically
pure form, on a gram scale in the case of the most potent
inhibitor, syn,syn-4. The variation of inhibitor potency as a
function of inhibitor stereochemistry was evaluated and proved
not to be straightforward, as discussed in detail.
Results and Discussion
The proposed structures (4) raised interesting questions about
the influence of the stereochemistry of the fluorinated center
upon enzyme binding. In a first-order analysis, fluorine substitu-
tion of the hydroxyethylene dipeptide isostere of indinavir is
expected to lower the pKa of the hydroxyl group and thereby
perhaps create a better mimic of the hydrated amide that is
implicated in peptide bond cleavage.20,21 The analysis quickly
becomes complex, however, when factors are considered such
as the differential solvation of free and enzyme-complexed
inhibitors, the influence of fluorination upon conformational
populations, potential steric and electronic interactions of the
enzyme and fluorine atom, and the variation of these factors as
a function of stereochemical changes at the fluorinated center.
Ultimately, it was clear that the only meaningful analysis would
be one based upon experimental findings.
Retrosynthetic Analysis. In analyzing routes to the targeted
inhibitors 4 it was considered that the fluorinated stereogenic
center, when coupled with the adjacent branched alkyl center
of asymmetry, presented the most challenging aspect of the
synthesis. The remainder of the molecule can be regarded as a
largely solved problem in light of synthetic studies that have
enabled the annual production of indinavir on the multi-ton
scale,22 work that has defined practical solutions for the
asymmetric synthesis of (S)-2-tert-butylcarboxamide-4-tert-
butoxycarbonyl-piperazine (6)23 and (1S,2R)-1-aminoindan-2-
ol (8)24 and that has established the carbon-nitrogen bonds
shown in Scheme 1 as convenient points of fragment assembly.
By this analysis, the synthesis of the inhibitor anti,anti-4, for
example, is reduced to the stereoselective construction of the
fluorinated core structure anti,anti-7, or its synthetic equivalent.
In retrosynthetic analysis of each of the core structures 7,
disconnection of the σ-bond that links the central stereo-
genic centers was an appealing and powerfully simplifying
strategy. A primary synthetic goal, therefore, became the
development of a method for the construction of this carbon-
carbon bond, one that would allow for the simultaneous and
selective formation of both stereogenic centers. In evaluating
σ-bond-forming strategies for the formation of this strategic
carbon-carbon bond, we considered applying asymmetric
alkylation methodology. In theory, either half of the targets 7
can be viewed as evolving from enolate- or electrophile-derived
In this work, we describe two successful approaches to the
synthesis of each of the four targeted monofluorinated HIV
(17) (a) Vacca, J. P.; Dorsey, B. D.; Schleif, W. A.; Levin, R. B.;
McDaniel, S. L.; Darke, P. L.; Zugay, J.; Quintero, J. C.; Blahy, O. M.;
Roth, E.; Sardana, V. V.; Schlabach, A. J.; Graham, P. I.; Condra, J. H.;
Gotlib, L.; Holloway, M. K.; Lin, J.; Chen, I.-W.; Vastag, K.; Ostovic, D.;
Anderson, P. S.; Emini, E. A.; Huff, J. R. Proc. Natl. Acad. Sci. U.S.A.
1994, 91, 4096-4100. (b) Dorsey, B. D.; Levin, R. B.; McDaniel, S. L.;
Vacca, J. P.; Guare, J. P.; Darke, P. L.; Zugay, J. A.; Emini, E. A.; Schleif,
W. A.; Quintero, J. C.; Lin, J. H.; Chen, I.-W.; Holloway, M. K.; Fitzgerald,
P. M. D.; Axel, M. G.; Ostovic, D.; Anderson, P. S.; Huff, J. R. J. Med.
Chem. 1994, 37, 3443-3451.
(18) The stereochemical descriptors given in compound labels in this
document refer to the stereochemical relationships in the central, variable
stereochemical triad. The first descriptor indicates the relative configuration
of the hydroxyl and fluorine substituents, while the second descriptor depicts
the relationship of the fluorine and benzyl substituents.
(19) Analogues of indinavir containing the monofluoro ketomethylene
dipeptide isostere were also targeted, see: Myers, A. G.; Barbay, J. K. Org.
Lett. 2001, 3, 425-428.
(20) For a review on the mechanism of peptide hydrolysis catalyzed by
HIV protease, see: Meek, T. D.; Rodriguez, E. J.; Angeles, T. S. In
RetroViral Proteases; Kuo, L. C., Shafer, J. A., Eds.; Methods in
Enzymology, Volume 241; Academic Press: New York, 1994; pp 127-
156.
(22) For a discussion on the industrial process for the asymmetric
synthesis of indinavir, see: Reider, P. J. Chimia 1997, 51, 306-308.
(23) (a) Askin, D.; Reider, P.; Rossen, K.; Varsolona, R. J.; Volante, R.
P.; Wells, K. M. Process for Making HIV Protease Inhibitors. U.S. Patent
5,618,939, April 8, 1997. (b) Askin, D.; Eng, K. K.; Rossen, K.; Purick, R.
M.; Wells, K. M.; Volante, R. P.; Reider, P. J. Tetrahedron Lett. 1994, 35,
673-676. (c) Rossen, K.; Weissman, S. A.; Sager, J.; Reamer, R. A.; Askin,
D.; Volante, R. P.; Reider, P. J. Tetrahedron Lett. 1995, 36, 6419-6422.
(d) Rossen, K.; Pye, P. J.; DiMichele, L. M.; Volante, R. P.; Reider, P. J.
Tetrahedron Lett. 1998, 39, 6823-6826.
(24) (a) Senanayake, C. H.; Roberts, F. E.; DiMichele, L. M.; Ryan, K.
M.; Liu, J.; Frendenburgh, L. E.; Foster, B. S.; Douglas, A. W.; Larsen, R.
D.; Verhoeven, T. R.; Reider, P. J. Tetrahedron Lett. 1995, 36, 3993-
3996. (b) Senanayake, C. H.; Smith, G. B.; Ryan, K. M.; Fedenburgh, L.
E.; Liu, J.; Roberts, F. E.; Hughes, D. L.; Larsen, R. D.; Verhoeven, T. R.;
Reider, P. J. Tetrahedron Lett. 1996, 37, 3271-3274. (c) Larrow, J. F.;
Roberts, E.; Verhoeven, T. R.; Ryan, K. M.; Senanayke, C. H.; Reider, P.
J.; Jacobsen, E. N. Org. Synth. 1999, 76, 46-56.
(21) For reviews on X-ray structures of HIV protease-inhibitor com-
plexes, see: (a) Wlodawer, A.; Erikson, J. W. Annu. ReV. Biochem. 1993,
62, 543-585. (b) Ringe, D. In RetroViral Proteases; Kuo, L. C., Shafer, J.
A., Eds.; Methods in Enzymology, Volume 241; Academic Press: New
York, 1994; pp 157-177. (c) Appelt, K. Perspect. Drug DiscoVery Des.
1993, 1, 23-48. (d) Babine, R. E.; Bender, S. L. Chem. ReV. 1997, 97,
1359-1472.

7210 J. Am. Chem. Soc., Vol. 123, No. 30, 2001
Myers et al.
Scheme 1
Addition of the nitroalkene 1226 (1.2 equiv) to a cold (-78
°C) solution of the lithium enolate derived from either enanti-
omer of pseudoephedrine R-fluoroacetamide resulted in the rapid
formation of conjugate addition products (Scheme 3). Spectro-
scopic analysis (1H, 19F NMR) of the crude reaction mixtures
showed that just two stereoisomers had been formed in each
reaction. It was later established that these isomers differed only
in the configuration of the â-center (ratio anti:syn ) 1.7:1); the
R-center had been formed with a single configuration, with the
same stereochemical preference as that found in pseudoephe-
drine enolate alkylation reactions using alkyl halides as
electrophiles.15d The nitroalkene that was recovered from the
product mixtures had undergone complete isomerization to the
corresponding styrene, indicating that deprotonation of the
electrophile competed with Michael addition. Nevertheless, the
reaction was amenable to relatively large-scale implementation;
on a 60-mmol scale, 15 g (65%) of the mixture of diastereomeric
adducts 11 and 13 was obtained. Synthetically useful quantities
of diastereomerically pure adduct 11 were obtained by recrys-
tallization of the mixture of diastereomeric products, typically
in 5-g batches (23% isolated yield).27
Scheme 2
Hydrolysis of the amide 11 occurred smoothly at 75 °C in a
mixture of 2 N aqueous sodium hydroxide, tert-butyl alcohol,
and methanol (Scheme 4). Following protonation of the
intermediate nitronate anion by the addition of glacial acetic
acid,28 the acid 14 was obtained in 98% yield (unpurified),
without detectable epimerization of the fluorinated center.
Conversion of the carboxylic acid 14 to the bromomethyl ketone
16 was accomplished in a two-step sequence involving forma-
tion of the diazomethyl ketone 15 (oxalyl chloride, catalytic
DMF, CH2Cl2; trimethylsilyldiazomethane, THF), followed by
treatment of this intermediate with hydrobromic acid.
Reduction of bromo ketone 16 with sodium borohydride
(EtOH, -78 °C) afforded the anti-fluorohydrin anti,anti-17
selectively, in 75% yield after chromatographic isolation (di-
astereoselectivity 9.4:1). The stereoselectivity of the reduction
is interesting and merits brief comment. The Anh-Eisenstein
model for nucleophilic addition to carbonyl compounds with
an adjacent polar group would have predicted the opposite
stereochemical outcome; this model invokes a Felkin-type
transition structure29 with the strongest acceptor ligand (in this
case fluorine) oriented antiperiplanar to the forming bond
(Scheme 4).30,31 Thus, in the case of R-fluoro ketones, this model
predicts preferential formation of syn-fluorohydrins, which was
not observed. A careful examination of reductions of R-fluoro
ketones reported in the literature,32 as well as of the data
presented herein, fails to reveal any compelling evidence to
support the dominance of an Anh-Eisenstein transition structure
in the reduction of R-fluoro ketones and suggests that the
stereochemical outcome is strongly dependent upon the reductant
used and upon subtle features of the substrate. In practice,
components and, in practice, both strategies were pursued with
success, this representing the fundamental distinction between
the two routes described herein.
Synthesis of the Targeted Protease Inhibitors Based on
Alkylation of Pseudoephedrine Fluoroacetamide. Initial
investigations focused on construction of the vicinal stereo-
centers of the fluorinated core structure by the alkylation of
pseudoephedrine R-fluoroacetamide (9)15a,25 with a variety of
electrophiles. The most direct route to the target structures using
this chiral fluoroacetamide enolate would involve alkylation with
a chiral, nonracemic secondary electrophile. A number of
secondary electrophiles bearing R-activating groups were ex-
plored as potential alkylation substrates; unfortunately, in no
circumstance were alkylation products isolated in yields greater
than 6%, as a consequence of the poor reactivity of secondary
electrophiles and the limited stability of this enolate above -40
°C. As an alternative strategy, an asymmetric Michael addition
reaction was considered for the formation of the key σ-bond.
The nitro compound anti,anti-10 was envisioned as a potential
precursor to the core structure anti,anti-7 by application of an
oxidative Nef transform (converting the carboxylic acid into a
nitro group in the retrosynthetic direction, Scheme 2). The nitro
compound anti,anti-10 was considered to arise, ultimately, from
the conjugate addition of the enolate derived from pseudoephe-
drine R-fluoroacetamide to the nitroalkene 12.
(26) (a) Ono, N.; Miyake, H.; Kamimura, A.; Kaji, A. J. Chem. Soc.,
Perkin Trans. 1 1987, 1929-1935. (b) Hass, H. B.; Susie, A. G.; Heider,
R. L. J. Org. Chem. 1950, 15, 8-14.
(27) The stereochemistry of ent-11 was established by X-ray crystal-
lographic analysisssee the Supporting Information.
(28) Acetic acid has been demonstrated to be an effective reagent for
the regeneration of nitroalkanes from their nitronate salts without significant
competing Nef reaction: Kornblum, N.; Graham, G. E. J. Am. Chem. Soc.
1951, 73, 4041-4043.
(29) (a) Che´rest, M.; Felkin, H.; Prudent, N. Tetrahedron Lett. 1968,
18, 2199-2204. (b) Che´rest, M.; Felkin, H. Tetrahedron Lett. 1968, 18,
2205-2208.
(30) Anh, N. T.; Eisenstein, O. NouV. J. Chem. 1977, 1, 61-70.
(31) For computational studies on the diastereoselectivity of the addition
of nucleophiles to R-fluoro ketones, see: (a) Wong, S. S.; Paddon-Row,
M. N. J. Chem. Soc., Chem. Commun. 1990, 456-458. (b) Wong, S. S.;
Paddon-Row, M. N. J. Chem. Soc., Chem. Commun. 1991, 327-330.
(25) CAUTION: Fluoroacetyl chloride, fluoroacetic acid, and derivatives
of fluoroacetic acid are exceedingly toxic, causing convulsions and
ventricular fibrillation upon inhalation and should be used only under
adequate supervision and in an appropriate fume hood. Although the specific
toxicities of 9 and other fluoroacetate derivatives described herein are
unknown, we urge that extreme caution be exercised in their preparation
and handling.

Synthesis of HIV Protease Inhibitors
J. Am. Chem. Soc., Vol. 123, No. 30, 2001 7211
Scheme 3
Scheme 4
Scheme 5
smoothly upon sequential treatment with potassium tert-butoxide
and potassium permanganate in a buffered mixture of tert-
butanol and water.33 The crude acid was coupled with (1S,2R)-
1-aminoindan-2-ol (8), affording amide anti,anti-19 in 89%
yield for the two-step sequence. Following cleavage of the
tetrahydropyranyl ether, the hydroxyamide within the resultant
diol was protected selectively in 93% yield by treatment with
2-methoxypropene and catalytic methanesulfonic acid. Rapid
and efficient epoxide formation then occurred upon treatment
of bromohydrin anti,anti-21 with potassium tert-butoxide in
THF at 0 °C.
Epoxide anti,anti-22 proved to be a convenient precursor to
two of the targeted indinavir analogues. Piperazine 25 (1.2 equiv,
prepared in two steps from (S)-1,2,4-piperazine tricarboxylic
acid, 1-tert-butyl-4-benzyl ester (23)34 as shown in Scheme 6)
reacted smoothly with this epoxide (1 equiv) upon heating at
80 °C in 2-propanol, in analogy with the commercial synthesis
of indinavir. Transfer hydrogenolysis of the benzyl carbamate
(1,4-cyclohexadiene, 10% Pd/C, EtOH) afforded the secondary
conditions for the selective reduction of a variety of R-fluoro
ketone substrates to either syn- or anti-configured fluorohydrin
products have been found by rapidly screening a variety of
simple, achiral reductants. A complete summary of diastereo-
selectivities observed in the reduction of R-fluoro ketones in
the course of this work is presented in graphical form in the
Supporting Information.
Protection of fluorohydrin anti,anti-17 as the corresponding
tetrahydropyranyl ether set the stage for application of the
oxidative Nef reaction (Scheme 5). Conversion of nitroalkane
anti,anti-18 to the corresponding carboxylic acid occurred
(32) For reductions of R-fluoro ketones in which the only stereogenic
center is the fluorinated carbon, see: (a) Barnett, J. E. G.; Kent, P. W. J.
Chem. Soc. 1963, 2743-2747. (b) Salon, M. C.; Hamman, S.; Beguin, C.
G. J. Fluorine Chem. 1985, 27, 361-370. (c) Brandange, S.; Dahlman, O.;
Olund, J. Acta Chem. Scand., Ser. B 1983, 37, 141-145. (d) Kitazume, T.;
Kobayashi, T.; Yamamoto, T.; Yamazaki, T. J. Org. Chem. 1987, 52, 3218-
3223. For reductions of R-fluoro ketones containing more than one
stereogenic center, see: (e) Bravo, P.; Piovosi, E.; Resnati, G. J. Chem.
Res. (S) 1989, 134-135; (M), 1115-1147. (f) Bravo, P.; Resnati, G.
Tetrahedron Lett. 1987, 28, 4865-4866. (g) Bravo, P.; Ganazzoli, F.;
Resnati, G.; De Munari, S.; Albinati, A. J. Chem. Res (S) 1988, 1701; (M),
1701-1736. (h) Ishihara, T.; Yamaguchi, K.; Kuroboshi, M.; Utimoto, K.
Tetrahedron Lett. 1994, 35, 5263-5266.
(33) (a) Shechter, H.; Williams, F. T., Jr. J. Org. Chem. 1962, 27, 3699-
3701. (b) Saville-Stones, E. A.; Lindell, S. D. Synlett 1991, 591-592. (c)
Go¨rlitzer, K.; Bo¨meke, M. Arch. Pharm. (Weinheim, Ger.) 1991, 324, 983-
987.
(34) Warshawsky, A. M.; Patel, M. V.; Chen, T.-M. J. Org. Chem. 1997,
62, 6439-6440.

7212 J. Am. Chem. Soc., Vol. 123, No. 30, 2001
Myers et al.
Scheme 6
Scheme 8
Scheme 7
pure fluorohydrin was isolated in 78% yield for the two-step
oxidation-reduction sequence. Acetonide deprotection afforded
fluorinated C17-epi-indinavir analogue syn,anti-4.
Having completed the synthesis of two of the targeted HIV
protease inhibitors, the problem of accessing compounds syn,-
syn-4 and anti,syn-4, having (R)-configuration at the fluorinated
center, remained to be addressed. These compounds were
prepared from Michael adduct ent-13 by synthetic sequences
closely related to those described above, for which complete
details are provided in the Supporting Information. Thus,
asymmetric alkylation of pseudoephedrine fluoroacetamide
enolate with a nitroalkene electrophile was successful in
providing access to the four stereoisomeric targets. However,
the moderate yield of the key Michael addition reaction
encouraged us to investigate alternative strategies for construc-
tion of the fluorinated core structures.
Two alternative, and ultimately unsuccessful, approaches
toward the key bond formation are summarized in Scheme 8,
with complete details provided in the Supporting Information.
The Michael addition of the lithium enolate derived from
pseudoephedrine fluoroacetamide to vinyl sulfoxide 30 produced
a single stereoisomeric adduct, but the efficiency of the reaction
was reduced by competing deprotonation of the electrophile.
Attempted Claisen rearrangement of secondary fluoroacetates
32 revealed complications due to elimination of fluoroacetic
acid from these substrates.37 These efforts were abandoned upon
discovery of an efficient and stereoselective means of synthesiz-
ing the fluorinated core subunit, based on an asymmetric
alkylation reaction employing nonracemic fluoroiodoacetic acid
as the electrophile, as described in detail below.
amine anti,anti-27 in 94% yield. Finally, alkylation of the
secondary amine with 3-picolyl chloride hydrochloride and
acidic methanolysis of the acetonide protecting group furnished
anti,anti-4, the first of the targeted protease inhibitors to be
prepared.
An oxidation-reduction strategy was employed to prepare
small quantities of inhibitor syn,anti-4 for determination of its
potential as an inhibitor of HIV protease (Scheme 7).35 Swern
oxidation of alcohol anti,anti-28 cleanly afforded ketone 29.36
Unexpectedly, ketone 29 was found to be unstable; decomposi-
tion of this compound occurred upon storage in solution (CH2-
Cl2) or during chromatography. We have found that R-amino
R′-fluoro ketones are generally unstable and, in a separate study,
have proposed and presented evidence for the facile formation
of a reactive oxyvinyliminium ion intermediate during their
decomposition.19 Ketone 29 was therefore reduced immediately
(LiBH(s-Bu)3, THF, -78 °C) without purification or storage.
Analysis of the crude reduction product by 19F NMR indicated
formation of syn-fluorohydrin syn,anti-28 without detectable
contamination by the anti diastereomer; the diastereomerically
Synthesis of the Targeted Protease Inhibitors Using
Nonracemic Fluoroiodoacetic Acid. Having encountered dif-
ficulties in the stereoselective construction of the fluorinated
core subunits by alkylation of a chiral fluoroacetamide enolate
or by fluoroacetate Claisen rearrangement, we became interested
in exploring an alternative asymmetric alkylation strategy
involving reversal of the polarity of the reaction components,
(35) Attempted inversion of alcohol anti,anti-28 under Mitsunobu
conditions (4-nitrobenzoic acid, Ph3P, DEAD, THF) failed due to a lack of
reactivity of this substrate.
(36) Mancuso, A. J.; Swern, D. Synthesis 1981, 165-185.
(37) Araki, K.; Welch, J. T. Tetrahedron Lett. 1993, 34, 2251-2254.

Synthesis of HIV Protease Inhibitors
J. Am. Chem. Soc., Vol. 123, No. 30, 2001 7213
Scheme 9
Scheme 10
have typically prepared this compound from the cheaper reagent
ethyl bromofluoroacetate by means of a Finkelstein reaction.
Crude ethyl fluoroiodoacetate produced in this manner was
subjected to basic hydrolysis (2 N aq LiOH, 0 °C) followed by
an acidic workup to provide fluoroiodoacetic acid in 86% overall
yield. This simple procedure required no chromatographic
purification and could be implemented without difficulty to
prepare >20 g of racemic fluoroiodoacetic acid in a single batch.
transferring the fluorine atom to the electrophilic partner.
Specifically, we proposed to use a chiral, nonracemic geminal
fluoroiodide as the electrophilic component. Issues that would
need to be addressed for the successful implementation of this
proposal included the identification of a suitably reactive, yet
stereochemically stable, electrophile that would undergo ste-
reospecific SN2 displacement, and the development of a practical
method for the asymmetric synthesis of such a species. With
regard to the issue of reactivity, kinetic studies have demon-
strated that R-fluorine substitution retards the rate of SN2
displacement of alkyl halides.38 The scarcity of methodology
for the asymmetric synthesis of geminal fluoroiodides presented
perhaps a greater obstacle. To the best of our knowledge, Bailey
and co-workers have reported the only previous syntheses of
optically pure R-fluorinated iodides, by chromatographic separa-
tion of the diastereomeric fluoroiodoacetamides derived from
R-methylbenzylamine.39 Importantly, displacement of iodide in
these compounds with nitrogen nucleophiles was demonstrated
to occur stereospecifically, with 85-100% inversion of stere-
ochemistry, providing precedent for the use of optically pure
geminal fluorohalides as reagents for the asymmetric synthesis
of organofluorine compounds.
Fluoroiodoacetic acid was ultimately selected as the electro-
phile in the proposed alkylation reaction after preliminary studies
with the alternative electrophiles 1,1-bromofluoro-2-tert-bu-
tyldimethylsilyloxyethane and tert-butyl fluoroiodoacetate had
shown the former to be unreactive and the latter to suffer from
competitive addition of nucleophiles to the ester carbonyl group.
These difficulties were avoided with fluoroiodoacetic acid, as
demonstrated by the alkylation of the enolate derived from
pseudoephedrine hydrocinnamide (35).15d Using racemic flu-
oroiodoacetic acid, the diastereomeric alkylation products 36
and 37 were formed in 80% yield (36:37 ) 1.2:1, Scheme 9).
The efficiency of this carbon-carbon bond-forming process
encouraged us to develop a procedure for the resolution of
fluoroiodoacetic acid, with the goal of obtaining a practical
solution for the preparation of either enantiomer on a multigram
scale.
Screening a variety of chiral bases for the resolution of
fluoroiodoacetic acid led to the identification of ephedrine as
the optimal resolving agent.42 Using this readily available and
inexpensive chiral reagent, either enantiomer of fluoroiodoacetic
acid can be prepared in g96% ee (Scheme 10). The optimized
procedure for the resolution involves heating a mixture of
racemic fluoroiodoacetic acid and (+)-ephedrine (hemihydrate)
or (-)-ephedrine in 2-propanol-methyl tert-butyl ether (9:1 v/v)
to reflux for 2 h.43 Epimerization of the fluorinated center occurs
under these conditions, allowing for dynamic resolution. Thus,
in theory, it is possible to obtain a quantitative yield of
diastereomerically pure product using racemic starting material.
In practice, the yields of crystalline material (80 and 84% de)
were 71 and 70% in resolutions using (-)-ephedrine and (+)-
ephedrine hemihydrate, respectively. Further diastereomeric
enrichment was achieved by successive recrystallizations from
ether-MeOH (10:1 v/v). Typically, four successive recrystal-
lization steps were required to obtain salt of g96% de. From a
practical standpoint, it is noteworthy that precipitation of the
salt was invariably rapid, allowing each recrystallization to be
conducted within 30 min; the entire resolution procedure can
be completed in a period of 7 h. From the diastereomerically
pure salt, enantiomerically enriched fluoroiodoacetic acid was
obtained in nearly quantitative yield and with e1% loss of
stereochemical purity by protonation with aqueous sulfuric acid.
The yields of enantiomerically enriched (R)- and (S)-fluor-
oiodoacetic acid were 43 and 39%, respectively, based on
racemic fluoroiodoacetic acid.44 This resolution procedure has
proven to be highly reproducible, routinely providing access to
highly enantiomerically enriched (g96% ee) fluoroiodoacetic
acid on scales up to 4.5 g. The enantiomerically enriched acid
is a fine, white powder that does not measurably epimerize upon
Fluoroiodoacetic acid can be readily prepared by basic
hydrolysis of ethyl fluoroiodoacetate.40 Although ethyl fluor-
oiodoacetate is now commercially available,41 in practice we
(41) Apollo Scientific Ltd., Derbyshire, UK.
(42) Other bases screened for the resolution included brucine, cinchoni-
dine, cinchonine, dehydroabietylamine, norephedrine, R-methylbenzylamine,
pseudoephedrine, quinidine, quinine, strychnine, and tyrosine ethyl ester.
(43) The use of enantiomeric resolving agents in different hydration states
(i.e., anhydrous (-)-ephedrine versus (+)-ephedrine hemihydrate) is a result
of commercial availability and did not cause significant differences in the
resolutions.
(44) The absolute stereochemistry of the enantiomers of fluoroiodoacetic
acid was determined by separate carbodiimide-mediated coupling reactions
of (R)-fluoroiodoacetic acid with each enantiomer of R-methylbenzylamine,
forming diastereomeric amides displaying physical properties (1H NMR,
19F NMR, melting point, TLC Rf values) in agreement with the values
reported in ref 39b.
(38) (a) McBee, E. T.; Christman, D. L.; Johnson, R. W.; Roberts, C.
W. J. Am. Chem. Soc. 1956, 78, 4595-4596. (b) Hine, J.; Thomas, C. H.;
Ehrenson, S. J. J. Am. Chem. Soc. 1955, 77, 3886-3887.
(39) (a) Bailey, P. D.; Boa, A. N.; Crofts, G. A.; van Diepen, M.;
Helliwell, M.; Gammon, R. E.; Harrison, M. J. Tetrahedron Lett. 1989, 30,
7457-7460. (b) Bailey, P. D.; Baker, S. R.; Boa, A. N.; Clayson, J.; Rosair,
G. M. Tetrahedron Lett. 1998, 39, 7755-7758.
(40) (a) Takeuchi, Y.; Takagi, K.; Yamaba, T.; Nabetani, M.; Koizumi,
T. J. Fluorine Chem. 1994, 68, 149-154. (b) Barth, F.; O-Yang, C.
Tetrahedron Lett. 1990, 31, 1121-1124.

7214 J. Am. Chem. Soc., Vol. 123, No. 30, 2001
Myers et al.
Scheme 11
Scheme 12
Scheme 13
prolonged storage at -20 °C in the dark.45,46 The chemical and
stereochemical stability of this compound significantly enhance
its potential as a useful two-carbon building block for the
asymmetric synthesis of organofluorine compounds.
Alkylations of pseudoephedrine hydrocinnamide (35) with
either enantiomer of fluoroiodoacetic acid proceeded in ap-
proximately 90% yield, with a high level of stereocontrol
(Scheme 11). In each alkylation reaction, each of the two newly
formed stereogenic centers was formed with >95% selectivity,
indicative of a high degree of stereospecificity in the displace-
ment reaction, as well as a high degree of diastereofacial
selectivity by the enolate, a characteristic of pseudoephedrine
amide enolate alkylations. These reactions provided an efficient
process for the formation of each of the diastereomeric
fluorinated core structures with high levels of stereocontrol and
could, in principle, form the basis of a synthetic route to the
targeted inhibitors. However, the availability of optically pure
fluoroiodoacetic acid and its now demonstrated participation
in stereospecific displacement reactions with amide enolate
nucleophiles, coupled with literature precedent for highly
diastereoselective alkylations of the enolate of aminoindanol
hydrocinnamide 40,47 raised the possibility of an even more
rapid and efficient route for the synthesis of inhibitors 4. This
strategy is depicted retrosynthetically in Scheme 12.
The crystalline amide 40 was prepared from commercial
(1S,2R)-1-aminoindan-2-ol (8) in 93% yield and without
chromatographic purification, following the literature proce-
dure.48 Conditions for the stereoselective alkylation of the
enolate derived from amide 40 with optically pure fluor-
oiodoacetic acid were developed. The optimized alkylation
procedure involved initial enolate formation (2.2 equiv amide
40, 2.3 equiv n-BuLi, THF, -78 °C) followed by addition of
(R)- or (S)-fluoroiodoacetic acid (1 equiv) and incubation at -60
°C. Excess base was quenched by the addition of water;
extractive isolation then provided the alkylation products 39 or
42 in >90% yield (see Scheme 13). The alkylation reaction
employing (R)-fluoroiodoacetic acid is a “matched” case; the
product 39 was obtained as a single diastereomer, within the
limits of detection by 19F NMR spectroscopy. The crude product
was typically converted directly to the methyl ester by treatment
with ethereal diazomethane, affording 41 in 82% yield for the
two-step sequence. “Mismatched” alkylation of 40, using (S)-
fluoroiodoacetic acid as the electrophile, afforded a 5.8:1 mixture
of epimers at the amide-derived center. A single recrystallization
of the crude reaction mixture from benzene afforded the
stereoisomerically pure alkylation product 42 in 70% yield. No
compounds epimeric at the fluorinated stereogenic center were
detected in either alkylation reaction, indicating that each
displacement reaction occurred with a high degree of ste-
reospecificity. The stereochemistry of the alkylation products
39 and 42 was determined unambiguously (see Supporting
Information), demonstrating that the displacement reactions
occur with inversion of stereochemistry at the fluorinated
stereogenic center. The inefficiency associated with the use of
an extra equivalent of the amide enolate to neutralize the acidic
proton of fluoroiodoacetic acid is to some degree offset by the
fact that the excess starting amide is recoverable (69-79% of
theory) and by the simplicity of the alkylation procedure. We
have used this method to prepare multigram quantities of
(45) A sample of acid (S)-38 having an enantiomeric excess of 99.8%
was stored in the dark at -20 °C for six months with no detectable loss of
stereochemical purity and without detectable decomposition.
(46) Fluoroiodoacetic acid is hygroscopic and was therefore weighed
under an inert atmosphere.
(47) Askin, D.; Wallace, M. A.; Vacca, J. P.; Reamer, R. A.; Volante,
R. P.; Shinkai, I. J. Org. Chem. 1992, 57, 2771-2773.
(48) Askin, D.; Eng, K. K.; Maligres, P. E.; Reider, P. J.; Rossen, K.;
Volante, R. P.; Upadhyay, V.; Weissman, S. A. Process for Making an
Epoxide. U.S. Patent 5,728,840, March 17, 1998.

Synthesis of HIV Protease Inhibitors
J. Am. Chem. Soc., Vol. 123, No. 30, 2001 7215
diastereomerically pure acids 39 and 42 (the largest reactions
to date have produced 5.7 g of 39 and 3.6 g of 42).49
Preliminary evidence suggested that activated acyl donors
derived from acids 39 and 42 were unstable, undergoing
intramolecular cyclization by attack of the amide carbonyl group,
raising doubts about the prospects of methods involving
carboxylic acid activation to transform the alkylation products
to chloromethyl ketones. This potential obstacle was overcome
by the development of a direct transformation of the methyl
esters 41 and 43 to the chloromethyl ketones 44 and 45 using
chloromethyllithium.50,51 In the optimized procedure, n-BuLi
(1.2 equiv) was added to a rapidly stirring mixture of the methyl
ester (41 or 43, 1 equiv) and chloroiodomethane (1.25 equiv)
in THF at -78 °C, followed by warming to 23 °C for 10 min
prior to an acidic quench. The chloromethyl ketones 44 and 45
were obtained in 88 and 89% yields, respectively, after column
chromatography. Spectroscopic analysis of the crude products
(19F NMR) showed no evidence of epimerization of the
fluorinated stereogenic center during the course of these
transformations.
Scheme 14
Scheme 15
Elaboration of intermediates 44 and 45 to the targeted
inhibitors involved conversion of the chloromethyl ketones to
the corresponding epoxides. Toward this end, the reduction of
these ketones with a number of simple, achiral reducing agents
was investigated, resulting in the development of conditions for
the stereoselective synthesis of each of the four targeted
fluorohydrin products 46 with diastereoselectivities of 12:1 or
greater (Scheme 14). The diastereomeric products were in all
cases separable by chromatography on silica gel, allowing the
isolation of products that were diastereomerically pure within
the limits of detection by 19F NMR.
(49) Alkylation products 39 and 42 are invariably contaminated with a
bright yellow-colored impurity, ascribed to the formation of trace amounts
(e3%) of isobenzofulvene 50 in the alkylation reactions. This assignment
is based on the isolation of the methyl ester 51 in 3% yield (trimethylsi-
lyldiazomethane, benzene, MeOH, 23 °C). Formation of the isobenzofulvene
impurity is proposed to occur as a result of benzylic deprotonation of amide
40, followed by elimination of acetone, alkylation of the resulting lithiated
enamide, tautomerization, and elimination of hydrofluoric acid. The
isobenzofulvene impurity can be chromatographically separated; therefore,
its formation does not significantly detract from the utility of this alkylation
protocol.
The conversion of alcohols 46 to the targeted protease
inhibitors was accomplished in three steps, following literature
precedent (Scheme 15).23b Thus, addition of potassium tert-
butoxide to solutions of each diastereomeric alcohol 46 in THF
led to rapid and clean formation of the corresponding epoxide.
The epoxides (1 equiv) were individually heated with (S)-2-
tert-butylcarboxamide-4-tert-butoxycarbonylpiperazine23 (6, 1.05
equiv, g99% ee) in 2-propanol at 75 °C. Following epoxide
opening, 6 N aqueous HCl was added, to deprotect the
carbamate- and acetonide-protective groups, simultaneously. The
secondary amines 47 were each obtained in good yield (70-
96%) for the two-step sequence. Finally, reductive alkylation
of amines 47 with 3-pyridinecarboxaldehyde using sodium
triacetoxyborohydride as reductant afforded the targeted com-
pounds in good yields.
(50) In situ generation of ClCH2Li in the presence of carbonyl com-
pounds: Sadhu, K. M.; Matteson, D. S. Tetrahedron Lett. 1986, 27, 795-
798.
(51) Application of ClCH2Li to the synthesis of R,R′-dihalogenated
ketones: (a) Barluenga, J.; Llavona, L.; Concello´n, J. M. J. Chem. Soc.,
Perkin Trans. 1 1990, 417. (b) Barluenga, J.; Llavona, L.; Concello´n, J.
M.; Yus, M. J. Chem. Soc., Perkin Trans. 1 1991, 297-300.
Biological Evaluation. Inhibition constants (Ki) for indinavir
(3), each of the four diastereomeric analogues 4, and C17-epi-

7216 J. Am. Chem. Soc., Vol. 123, No. 30, 2001
Myers et al.
Table 1. Inhibition Constants (Ki) for Inhibition of HIV-1 Protease by Indinavir and Analogues
indinavir (48)52 were determined with respect to HIV-1 protease
using a standard assay wherein the initial rate of enzymatic
cleavage of the synthetic substrate Lys-Ala-Arg-Val-Nle-Nph-
Glu-Ala-Nle-NH2 is measured in the presence of varying
concentrations of inhibitor.53,54 Inhibition constants are presented
in Table 1. The most potent of the fluorinated inhibitors was
the diastereomer syn,syn-4; this compound was essentially
equipotent to indinavir in this assay. Inhibition constants varied
over a range of 3 orders of magnitude among the series of
diastereomers 4.
Although the data set is small, it is nevertheless apparent
that insertion of fluorine into the hydroxyethylene cores of
indinavir or C17-epi-indinavir to form syn-fluorohydrins has a
more favorable impact on potency than the alternative substitu-
tions, producing anti-configured fluorohydrins. Thus, syn-
fluorine substitution into the indinavir and epi-indinavir frame-
works results in an equipotent and an 8-fold more potent
inhibitor, respectively, while anti-fluorine substitution results
in 14- and 37-fold less potent inhibitors. Qualitative analysis
of proton-proton and proton-fluorine coupling constants
suggests that the syn-fluorohydrin-containing inhibitors syn,-
syn-4 and syn,anti-4 maintain a fully extended conformation
of the carbon-carbon chain encompassing the fluorohydrin
subunit (C16-C19). In contrast, the anti-fluorohydrin-containing
inhibitors anti,anti-4 and anti,syn-4 exhibit coupling constants
indicative of the presence of significant populations of alterna-
tive conformers. When bound to HIV protease, the carbon
backbone of indinavir assumes an extended conformation in the
region of the central hydroxyl (dihedral angles (C16-C17)-
(C18-C19) ) 178.6° and 173.2° when bound to HIV-1 and
HIV-2 proteases, respectively).57 The weaker binding of inhibi-
tors containing anti-fluorohydrin subunits, relative to nonflu-
orinated or syn-fluorohydrin inhibitors, may be due in part to
the larger population of solution conformations of these inhibi-
tors that differ significantly from the enzyme-bound conforma-
tion.
In an attempt to rationalize the effect of the stereochemistry
of the monofluoro hydroxyethylene core on the potency of
inhibitors 4, we examined proton-proton and proton-fluorine
coupling constants in these compounds and all fluorohydrin
synthetic intermediates employed in this study. A graphical
summary of the spectral characteristics of these fluorohydrins
is presented in Figure 2. Three patterns are invariant in the
series: (1) each anti-fluorohydrin exhibits a greater vicinal
CHOH-CHF proton-proton coupling constant than any of the
syn-fluorohydrins (Figure 2a), (2) each anti-fluorohydrin has a
smaller vicinal CHOH-CHF proton-fluorine coupling constant
than any of the syn-fluorohydrins (Figure 2b), and (3) within a
given structural framework, the anti-fluorohydrins display 19F
NMR resonances shifted downfield in comparison to the
corresponding syn-fluorohydrins (Figure 2c). All three trends
have been observed previously in other acyclic fluorohydrins.55
The observed trends in coupling constants can be used to gain
insight into the conformational preferences of fluorohydrins, as
vicinal proton-fluorine coupling constants exhibit a Karplus-
type dependence on dihedral angle.56 The observed vicinal
coupling constants for both syn- and anti-fluorohydrins are
qualitatively consistent with preferred solution conformations
wherein the carbon chain is fully extended (see Newman
projections, Figure 2), although it is apparent that the anti-
fluorohydrins display a greater degree of conformational vari-
ability than their syn counterparts.
The difference in conformational preferences of syn- and anti-
fluorohydrin inhibitors may be a consequence of the gauche
effect. The gauche effect is a tendency for vicinal electronegative
substituents to adopt a gauche conformation.58,59 The physical
(56) Caution must be exercised in this analysis, as it has been
demonstrated that proton-fluorine coupling constants are more sensitive
than proton-proton coupling constants to C-C-H and C-C-F bond
angles and to substituent effects, see: (a) Thibaudeau, C.; Plavec, J.;
Chattopadhyaya, J. J. Org. Chem. 1998, 63, 4967-4984 and references
therein. (b) Fabian, J. S.; Guilleme, J. Chem. Phys. 1996, 206, 325-337
and references therein. (c) Ihrig, A. M.; Smith, S. L. J. Am. Chem. Soc.
1970, 92, 759-763. (d) Williamson, K. L.; Hsu, Y.-F. L.; Hall, F. H.;
Swager, S.; Coulter, M. S. J. Am. Chem. Soc. 1968, 90, 6717-6722. (e)
Williamson, K. L.; Hsu, Y.-F. L.; Hall, F. H.; Swager, S. J. Am. Chem.
Soc. 1966, 88, 5678-5680. (f) Hamman, S.; Beguin, C.; Charlon, C.; Luu-
Duc, C. Org. Magn. Reson. 1983, 21, 361-366.
(57) X-ray structure of indinavir bound to HIV-2 protease: Chen, Z.;
Li, Y.; Chen, E.; Hall, D. L.; Darke, P. L.; Culberson, C.; Shafer, J.; Kuo,
L. C. J. Biol. Chem. 1994, 269, 26344-26348. The structure of indinavir
bound to HIV-1 protease is available electronically (National Cancer
Institute’s HIV protease database: www.ncifcrf.gov/CRYS/HIVdb, structure
“1hsg”).
(52) The preparation of C17-epi-indinavir (48) is described in the
Supporting Information.
(53) Abbreviations: Nle, norleucine; Nph, 4-nitrophenylalanine.
(54) (a) Richards, A. D.; Phylip, L. H.; Farmerie, W. G.; Scarborough,
P. E.; Alvarez, A.; Dunn, B. M.; Hirel, P.-H.; Konvalinka, J.; Strop, P.;
Pavlickova, L.; Kostka, V.; Kay, J. J. Biol. Chem. 1990, 265, 7733-7736.
(b) Phylip, L. H.; Richards, A. D.; Kay, J.; Konvalinka, J.; Strop, P.; Blaha,
I.; Velek, J.; Kostka, V.; Ritchie, A. J.; Broadhurst, A. V.; Farmerie, W.
G.; Scarborough, P. E.; Dunn, B. M. Biochem. Biophys. Res. Commun.
1990, 171, 439-444.
(55) Bravo, P.; Piovoisi, E.; Resnati, G. J. Chem. Res. (M) 1988, 1115-
1147 and references therein.
(58) Wolfe, S. Acc. Chem. Res. 1972, 5, 102-111.

Synthesis of HIV Protease Inhibitors
J. Am. Chem. Soc., Vol. 123, No. 30, 2001 7217
Figure 2. Comparison of NMR spectral characteristics of syn- and anti-fluorohydrins. Underlined compound numbers indicate data from methanol-
d4 solutions. All other data are from chloroform-d solutions.
basis for this conformational preference is thought to be a
stabilizing hyperconjugative interaction between a σ-bonding
orbital and an adjacent σ* antibonding orbital that is maximized
when σ-bonds having the greatest donor capacity (i.e., C-H)
are oriented antiperiplanar to the strongest acceptor σ bonds
(in this case, C-F). The gauche effect is expected to stabilize
the fully extended conformation of the carbon backbone for syn-
fluorohydrins, as this conformation orients C-H bonds anti-
periplanar to both electronegative substituents. In contrast, in
anti-fluorohydrins the electronegative substituents are oriented
(59) For leading references on the gauche effect, see: (a) Juaristi, E.;
Cuevas, G. Tetrahedron 1992, 48, 5019-5087. (b) Thatcher, G. R. J., Ed.
The Anomeric Effect and Associated Stereoelectronic Effects; ACS Sym-
posium Series 539; American Chemical Society: Washington, DC, 1993.

7218 J. Am. Chem. Soc., Vol. 123, No. 30, 2001
Myers et al.
Table 2. Properties of Indinavir and Analogs
in an electronically unfavorable antiperiplanar relationship in
the extended conformer.60
inhibition of
inhibition of
CYP2D6, IC50
(µM)
CYP3A4, IC50
Considering only the syn-fluorohydrin inhibitors, where
fluorination appears to reinforce the extended inhibitor confor-
mation observed in the structure of HIV protease bound to
indinavir, it is interesting to note that fluorination of indinavir
(3 f syn,syn-4) produces no change in potency, while fluorine
substitution in the core of C17-epi-indinavir (48 f syn,anti-4)
led to an 8-fold increase in potency. Aspartyl proteases display
consistently stronger binding to hydroxyethylene-containing
inhibitors in which the core hydroxyl has (S) absolute config-
uration (for instance, indinavir) relative to inhibitors epimeric
at this stereogenic center; the approximately 90-fold loss in
potency observed on inversion of this center in indinavir is
consistent with the differences observed in other inhibitors of
HIV protease.61 The central (S)-configured hydroxyl in potent
inhibitors is typically bound between the two catalytic aspartates,
occupying the site taken by the nucleophilic water molecule in
the structure of the native enzyme.21,62 Previous observations
showing that difluorostatine-containing inhibitors of the aspartyl
proteases renin and pepsin are less potent63 or essentially
equipotent9a in comparison to the corresponding nonfluorinated
inhibitors suggest that the lack of improved binding of syn,-
syn-4 compared to that of indinavir may be indicative of a
general trend. This lack of enhanced potency upon fluorination
has been previously suggested to reflect the importance of the
central alcohol as a hydrogen-bond acceptor.63a,b The tighter
binding of the enzyme to syn,anti-4, as compared to its
nonfluorinated counterpart (C17-epi-indinavir, 48), may be due
in part to formation of a stronger hydrogen bond between the
catalytic aspartate(s), which are expected to bear a net negative
charge,20 and the â-fluorinated core hydroxyl, due to its
increased hydrogen-bond donor ability. The lack of improved
potency upon fluorination of indinavir may be the result of a
different pattern of hydrogen bonding between the core hydroxyl
and the catalytic aspartates in inhibitors varying in the stereo-
chemistry of the central alcohol. Alternatively, it is conceivable
that hydrogen bonding to the catalytic aspartates is strengthened
for both syn,syn-4 and syn,anti-4, but the effect is diminished
by other, unfavorable, consequences of fluorination in the case
of syn,syn-4. Binding energy is critically dependent upon factors
such as solvation, hydrophobic and steric interactions with the
enzyme, and populations of alternative inhibitor conformations,
compound
indinavir (3)
C17-epi-indinavir (48)
syn,syn-4
anti,anti-4
anti,syn-4
log Pa
(µM)
3.03
3.02
3.23
3.01
3.31
3.12
1.7
no data
1.6
1.7
2.4
82.6
no data
78.9
57.2
35.3
syn,anti-4
7.5
198
a Log 1-octanol/water partition coefficient, determined by the shake-
flask method.64
each of which may be altered, in poorly defined ways, upon
introduction of fluorine.
The lipophilicity and the extent of inhibition of human
cytochrome P450 (CYP) by inhibitors 4 and indinavir were
determined (Table 2). Each of the fluorinated inhibitors was
found to be slightly more lipophilic (e0.28 log units) than
indinavir, with the exception of anti,anti-4, which had a
measured 1-octanol-water partition coefficient64 essentially
equal to that of the unaltered drug. These results are consistent
with the trend that fluorination of carbons adjacent to heteroatom
substituents results in increased lipophilicity.3a The lipophilicity
of drug candidates is an important determinant of pharmokinetic
factors such as absorption, cell membrane permeability, plasma
protein binding, metabolism, and penetration into the central
nervous system.65 Indinavir and its fluorinated analogues 4
displayed similair inhibition of cytochrome P450 isozymes
CYP3A4 and CYP2D6, enzymes responsible for metabolism
of indinavir.
Summary
We have developed syntheses of four diastereomeric HIV
protease inhibitors (4) incorporating the monofluoro hydroxy-
ethylene dipeptide isostere. The two successful routes presented
employed asymmetric alkylation reactions for stereoselective
construction of the fluorinated core structures. Studies involving
the use of a fluorinated nucleophile (pseudoephedrine R-fluo-
roacetamide) and various nonfluorinated electrophiles in the key
carbon-carbon bond-forming reaction revealed complications
introduced by the presence of the branched alkyl center of
asymmetry in the core structures (lack of reactivity in the case
of secondary electrophiles; benzylic deprotonation or lack of
diastereoselectivity with Michael acceptors). The alternative
carbon-carbon bond construction based on the alkylation of
an amide enolate with an R-fluorinated electrophile (enantio-
merically enriched fluoroiodoacetic acid) resulted in the devel-
opment of practical and efficient synthetic routes to the targeted
compounds. The optimized synthetic route to the most potent
inhibitor (syn,syn-4, Ki ) 2.0 nM) proceeded in seven steps
(87% average yield per step) from aminoindanol hydrocinna-
mide 40 and (S)-fluoroiodoacetic acid. A simple procedure for
the resolution of fluoroiodoacetic acid was developed, providing
access to either enantiomer in g96% ee. The alkylation of the
enolate derived from amide 40 with optically active fluor-
(60) For studies on the gauche effect in fluorohydrins, see: (a) Abraham,
R. J.; Chambers, E. J.; Thomas, W. A. J. Chem. Soc., Perkin Trans. 2 1994,
949-955. (b) Buckley, P.; Giguere, P. A.; Daijiro, Y. Can. J. Chem. 1968,
46, 2917-2923. (c) Griffith, R. C.; Roberts, J. D. Tetrahedron Lett. 1974,
39, 3499-3502. (d) Bakke, J. M.; Bjerkeseth, L. H.; Ronnow, T. E. C. L.;
Steisvoll, K. J. Mol. Struct. 1994, 321, 205-214. (e) Huang, J.; Hedberg,
K. J. Am. Chem. Soc. 1989, 111, 6909-6913. (f) Braathen, O.-A.; Marstokk,
K.-M.; Mollendal, H. Acta Chem. Scand. 1982, 6, 173-181. (g) Hagen,
K.; Hedberg, K. J. Am. Chem. Soc. 1973, 95, 8263-8266.
(61) (a) Dreyer, G. B.; Metcalf, B. W.; Tomaszek, T. A.; Carr, T. J.;
Chandler, A. C.; Hyland, L.; Fakhoury, S. A.; Magaard, V. W.; Moore, M.
L.; Strickler, J. E.; Debouck, C.; Meek, T. D. Proc. Natl. Acad. Sci. U.S.A.
1989, 86, 9752-9756. (b) Rich, D. H.; Sun, C.-G.; Prasad, J. V. N.;
Pathiasseril, A.; Toth, M. V.; Marshall, G. R.; Clare, M.; Mueller, R. A.;
Houseman, K. J. Med. Chem. 1991, 34, 1222-1228.
(62) For the native structure of HIV-1 protease, see: Navia, M. A.;
Fitzgerald, P. M. D.; McKeever, B. M.; Leu, C.-T.; Heimbach, J. C.; Herber,
W. K.; Sigal, I. S.; Darke, P. L.; Springer, J. P. Nature 1989, 337, 615-
620.
(63) (a) Thaisrivongs, S.; Pals, D. T.; Kati, W. M.; Turner, S. R.;
Thomasco, L. M.; Watt, W. J. Med. Chem. 1986, 29, 2080-2087. (b)
Fearon, K.; Spaltenstein, A.; Hopkins, P. B.; Gelb, M. H. J. Med. Chem.
1987, 30, 1617-1622. (c) Doherty, A. M.; Kaltenbronn, J. S.; Hudspeth, J.
P.; Repine, J. T.; Roark, W. H.; Sircar, I.; Tinney, F. J.; Connolly, C. J.;
Hodges, J. C.; Taylor, M. D.; Batley, B. L.; Ryan, M. J.; Essenberg, A. D.;
Rapundalo, S. T.; Weishar, R. E.; Humblet, C.; Lunney, E. A. J. Med. Chem.
1991, 34, 1258-1271.
(64) Leo, A.; Hansch, C.; Elkins, D. Chem. ReV. 1971, 71, 525-616.
(65) (a) Navia, M. N.; Chaturvedi, P. R. Drug DiscoVery Today 1996,
5, 179-189. For reviews on the pharmacokinetics of anti-HIV agents, see:
(b) Williams, G. C.; Sinko, P. J. AdV. Drug DeliVery ReV. 1999, 39, 211-
238. (c) Li, X.; Chan, W. K. AdV. Drug DeliVery ReV. 1999, 39, 81-103.
(d) Aungst, B. J. AdV. Drug DeliVery ReV. 1999, 39, 105-116. For a review
focusing on pharmacokinetic aspects of the development of indinavir, see:
(e) Lin, J. H.; Ostovic, D.; Vacca, J. P. In Integration of Pharmaceutical
DiscoVery and DeVelopment: Case Histories; Borchardt, R. T., Freidinger,
R. M., Sawyer, T. K., Smith, P. L., Eds. Pharmaceutical Biotechnology,
Volume 11; Plenum Press: New York, 1998; pp 233-255.

Synthesis of HIV Protease Inhibitors
J. Am. Chem. Soc., Vol. 123, No. 30, 2001 7219
oiodoacetic acid proceeded with a high degree of stereospeci-
ficity. Coupled with the development of methods for the
diastereoselective reduction of R-fluoro ketones 44 and 45, this
route allowed for the efficient and stereoselective synthesis of
inhibitors containing each of the four stereoisomeric forms of
the monofluoro hydroxyethylene dipeptide isostere, more than
one gram in the case of the inhibitor syn,syn-4. The inhibitory
potency and other relevant biophysical properties of these
structurally novel protease inhibitors were determined.
Enanta Pharmaceuticals for determination of the inhibition of
human cytochrome P450 by compounds 4.
Supporting Information Available: A description of first-
generation syntheses of inhibitors anti,syn-4 and syn,syn-4,
alternative approaches to the fluorinated core subunits, prepara-
tion of C17-epi-indinavir, and stereochemical assignments,
graphical summaries of diastereoselectivity in the reduction of
R-fluoro ketones, synthetic procedures, and characterization data
for all new compounds, and X-ray crystal data for compounds
ent-11 and anti,syn-22 (PDF). This material is available free
of charge via the Internet at http://pubs.acs.org.
Acknowledgment. Financial support from the National
Science Foundation and the National Institutes of Health is
gratefully acknowledged. We thank Dr. Michael Day and Dr.
Richard Staples for obtaining crystallographic data. We express
our appreciation to Dr. Jens Eckstein and Dr. Greg Warner of
JA010113Y