Novel prodrug approach to amprenavir-based HIV-1 protease inhibitors via O→N acyloxy migration of P1 moiety

Article abstract of DOI:10.1016/S0960-894X(03)00463-3

We have developed a new approach to prodrugs, which utilizes a pH-induced intramolecular O→N migration of an acyloxy group in carbonate moiety to a free amino moiety at neutral pH. This method is exemplified by facile rearrangement of highly water-soluble prodrug 3 to carbamate 4, a close analogue of HIV-1 protease inhibitor Amprenavir. The O→N acyloxy migration is unprecedented in the context of prodrugs and it enables a high atom economy due to recycling of the 'pro' moiety.

Full text of DOI:10.1016/S0960-894X(03)00463-3

Bioorganic & Medicinal Chemistry Letters 13 (2003) 2523–2526
Novel Prodrug Approach to Amprenavir-Based HIV-1 Protease
Inhibitors Via O!N Acyloxy Migration of P1 Moiety
Wieslaw M. Kazmierski,* Patricia Bevans, Eric Furfine,
Andrew Spaltenstein and Hanbiao Yang
GlaxoSmithKline, 5 Moore Drive, Research Triangle Park, NC 27709, USA
Received 30 March 2003; revised 25 April 2003; accepted 6 May 2003
Abstract—We have developed a new approach to prodrugs, which utilizes a pH-induced intramolecular O!N migration of an
acyloxy group in carbonate moiety to a free amino moiety at neutral pH. This method is exemplified by facile rearrangement of
highly water-soluble prodrug 3 to carbamate 4, a close analogue of HIV-1 protease inhibitor Amprenavir. The O!N acyloxy
migration is unprecedented in the context of prodrugs and it enables a high atom economy due to recycling of the ‘pro’ moiety.
# 2003 Elsevier Ltd. All rights reserved.
HIV-1 protease inhibitors (PI), such as Amprenavir 1,1
are key components in the highly active antiretroviral
therapy (HAART), which has been credited with
decreasing mortality rates from AIDS in the developed
world. However, a heavy pill burden of HAART some-
times results in a limited patient adherence to the drug
regimen, thus reducing or limiting the benefits of
HAART and allowing emergence of viral resistance,
rebound of viral load, and ultimately a disease progres-
sion. Much of the pill burden associated with PIs is the
result of their generally low aqueous solubility. Conse-
quently, there is a continued need to discover PIs with
superior potency, resistance profile, and with improved
solubility. Based on this premise, some of our research
has been directed towards novel and soluble prodrugs of
Amprenavir. We have recently described some of these
activities,2 which culminated in the discovery of phos-
phate 2,3ꢀ6 currently in advanced stages of clinical
development (Fig. 1).
Although conceptually related prodrug strategy, which
utilizes acyl transfer has been described,8,9 the O!N
acyloxy migration reported herein is novel in the con-
text of prodrugs. Extensive literature searches revealed
that reports of intramolecular O!N acyloxy migration
are rare and usually occur in aromatic systems such
as anilines under forcing, non-physiological con-
ditions.10ꢀ13 In spite of this limited precedent, we
decided to explore this transformation in the context
of Amprenavir prodrugs.
In order to avoid the necessary protection/deprotection
of the anilino group in Amprenavir, we based our
explorations on the nitro-derivative 3. Our initial strat-
egy towards 3 involved the synthesis of Boc-protected 7,
which could be obtained from intermediate 5, pre-
viously utilized in syntheses of other Amprenavir ana-
logues,3ꢀ6 and 3-tetrahydrofurane chloroformate. This
approach was first evaluated with commercially avail-
able methyl and allyl chloroformates. Low to medium
yields of respective carbonates 10 and 11 necessitated
time-consuming chromatographies,14 and in order to
optimize this chemistry we embarked on an alternative
route employing nitrophenyl carbonate 6. After a con-
siderable experimentation, we developed an efficient
synthesis of 6 from 5 and bis (p-nitro-phenyl) carbonate
in presence of either P4-phosphazene base (61% yield),
or lithium hydroxide (80% yield, Fig. 3). Next, target
molecule 715 was conveniently synthesized by reacting 6
with (S)-(+)-3-hydroxy-tetrahydrofuran in refluxing
dioxane. Other carbonates, including those of primary
In the course of our investigations we discovered that
the mixed carbonate 3 undergoes O!N acyloxy migra-
tion to 4, a nitro derivative of Amprenavir 1 (Fig. 2).
We now wish to report on this new class of water-solu-
ble prodrugs, exemplified by 3,3ꢀ6 which readily con-
verts to drug molecules neutral pH at room
temperature.7
*Corresponding author. Tel.: +1-919-483-9462; fax: +1-919-315-
0430; e-mail: wieslaw.m.kazmierski@gsk.com
0960-894X/03/$ - see front matter # 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/S0960-894X(03)00463-3

2524
W. M. Kazmierski et al. / Bioorg. Med. Chem. Lett. 13 (2003) 2523–2526
Figure 1. Chemical structures of Amprenavir 1 and its phosphate
prodrug 2.
Figure 4. Substituent effect on the kinetics of O!N acyloxy migration
at pH=7.0. Left to right, prodrugs 3, 10 and 11.
Figure 2. Structures of nitro-Amprenavir 4 and its carbonate prodrug
3.
Figure 5. Formation of oxazolidin-2-ones in the O!N acyloxy
migration.
Detailed analysis of this reaction revealed formation of
minor quantity of oxazolidin-2-one, presumably by
elimination of alcohol during the acyloxy rearrange-
ment. We further determined ratios of drug to oxazoli-
1
din-2-one by H NMR (Fig. 5). Allyl prodrug 11, the
slowest converting compound in this set (Fig. 4), gave
the highest ratio of drug to oxazolidin-2-one (Fig. 5).
We speculated that more stable prodrugs could further
reduce or even eliminate the formation of oxazolidin-2-
one.
.
Figure 3. (a) LiOH H2O in dioxane, bis-(p-nitrophenyl) carbonate;
(b) dioxane, 3 equiv triethylamine; (c) 50% TFA in DCM; (d) 1:1 (v/v)
buffer/THF.
In conclusion, we have developed efficient chemistry
towards novel mixed-carbonate prodrugs of the type
3, and demonstrated that under simulated physio-
logical conditions these undergo swift O!N acyloxy
migration to carbamate-based protease inhibitors,
such as 4. Such a migration is unprecedented in the
context of prodrugs and differs from the more classi-
cal hydrolytic approaches to prodrugs, in which the
‘pro’-moiety is hydrolized off. In these cases, the ‘pro’-
moiety is irretrievably lost, resulting in a poor atom
economy. In addition, potential toxicity issues of
cleaved ‘pro’-moieties limit their choices. We believe
that the approach described herein addresses some of
these drawbacks and that it can be utilized generally
as a prodrug approach towards carbamate moiety-
bearing drugs.
alcohols could also be obtained using similar metho-
dology (not shown). Subsequent TFA-mediated depro-
tection of Boc carbonate 7 cleanly yielded the TFA salt
of 3 (Fig. 3).
We then subjected 3 to simulated physiological condi-
tions [1:1 (v/v) buffer/THF at room temperature] and
attempted to induce the O!N acyloxy migration. We
were pleased to observe a fast conversion of the car-
bonate to new product, later found identical to the
authentic sample of 4.3,5,6 We also demonstrated that
analogous migration occurred in both methyl (10) and
allyl (11)14 prodrugs (Fig. 4), which pointed to a general
applicability of this approach.

W. M. Kazmierski et al. / Bioorg. Med. Chem. Lett. 13 (2003) 2523–2526
2525
References and Notes
3.61 (3H, s), 3.38 (1H, m), 3.26 (1H, m), 2.95 (1H, m), 2.89
(1H, m), 2.79 (2H, m), 1.79 (1H, m), 0.70 (6H, pseudo d).
O!N acyloxy migration: 1:1 (v/v) mixture of THF and
potassium phosphate monobasic–sodium hydroxide buffers
(0.05 M, pH 7.0 or 8.0) were used to investigate the O!N
migration. Concentration of prodrug in the buffer was ca. 6
mM. Thus, 10 (0.030 g, 0.0625 mmol) was dissolved in phos-
phate buffer (pH 7.0) and THF (5+5 mL), and the reaction
was LC/MS monitored at room temperature. Upon comple-
tion, preparative HPLC purification yielded 4 mg of 12. 1H
NMR (400 MHz, CDCl3) d 8.33 (2H, d, J=8.8 Hz), 8.02 (2H,
d, J=8.8 Hz), 7.21-7.11 (5H, m), 7.02 (1H, m), 5.02 (1H, m),
3.60 (1H, m), 3.41 (1H, m), 3.35 (3H, s), 3.02-2.96 (2H, m),
2.91 (1H, m), 2.86 (2H, m), 2.83 (1H, m), 1.91(1H, m), 0.82-
0.77 (6H, m). LRMS: m/z 480.1 (MH+).
1. Kim, E. E.; Baker, C. T.; Dwyer, M. D.; Murcko, M. A.;
Rao, B. G.; Tung, R. D.; Navia, M. A. J. Am. Chem. Soc.
1995, 117, 1181.
2. Kazmierski, W. M.; Furfine, E.; Spaltenstein, A.; Wright,
L. L. Bioorg. Med. Chem. Lett. 2002, 12, 3431.
3. Tung, R., Hale, M., Baker, C., Furfine, E.; Kaldor, I.;
Kazmierski, W., Spaltenstein, A. PTC Int. Appl. WO
9933815.
4. Hale, M.; Tung, R.; Baker, C.; Spaltenstein, A.; Furfine. E.;
Kaldor, I., Kazmierski, W. PCT Int. Appl. WO 9933795
5. Hale, M.; Tung, R.; Baker, C.; Spaltenstein, A.; Furfine, E.;
Kaldor, I., Kazmierski, W. PCT Int. Appl. WO 9933793
6. Hale, M.; Tung, R.; Baker, C.; Spaltenstein, A.; Furfine, E.;
Kaldor, Kazmierski, W. PCT Int. Appl. WO 9933792
7. Kazmierski, W. M.; Bevans, P.; Furfine, E.; Porter,
D.; Spaltenstein, A.; Yang, H. 223rd ACS National
Meeting, Orlando, FL, United States, April 7–11, 2002;
MEDI-160.
As expected, conversion rates generally were faster at pH 8
(as exemplified below).
8. Kiso, Y.; Matsumoto, H.; Yamaguchi, S.; Kimura, T. Lett.
Pept. Sci. 1999, 6, 275.
9. Hamada, Y.; Ohtake, J.; Sohma, Y.; Kimura, T.; Hayashi,
Y.; Kiso, Y. Bioorg. Med. Chem. 2002, 10, 4155.
10. Stieglitz, U. Am. Chem. J. 1904, 31, 501.
11. Ransom, A. Am. Chem. J. 1900, 23, 43.
12. Bernet, B.; Bishop, P. M.; Caron, M.; Kawamata, T.;
Roy, B. L.; Ruest, L.; Sauve, G.; Soucy, P.; Deslongchamps,
P. Can. J. Chem. 1985, 63, 2818.
13. Kato, S.; Morie, T. J. Heterocycl. Chem. 1996, 33, 1171.
14. 1.
15. The synthesis of 6
Phosphazene method. Bis-(p-nitrophenyl) carbonate (13.39 g,
43.56 mmol in 44.62 mL DMF) and phosphazene base P4-t-Bu
(10.89 mL, 1 M in hexane, 10.89 mmol) were added to (0.50 g,
0.99 mmol) of 5. The mixture was stirred at room temperature
for approximately 6 h, washed with ethyl acetate, 1 N HCl,
brine, and 1 N NaOH sequentially, dried over MgSO4, and
concentrated. The residue was purified by silica gel column
chromatography with ethyl acetate/hexane (1/2) (61% yield of
6).
1H NMR: d 8.41 (2H, d, J=8.8 Hz), 8.34 (2H, d, J=9.0 Hz),
8.11 (2H, d, J=8.8 Hz), 7.47 (2H, d, J=9.1 Hz), 7.30–7.23
(5H, m), 7.21 (1H, m), 4.94 (1H, m), 4.17 (1H, m), 3.59–3.45
(2H, m), 3.04–2.91 (2H, m), 2.84–2.67 (2H, m), 1.92 (1H, m),
1.29 (9H, s), 0.85 (6H, d, J=6.1 Hz). LRMS: m/z 687.3
(MH+).
LiOH method. LiOH monohydrate (0.089 g, 2.2 mmol) was
added to (0.10 g, 0.192 mmol) of 5 and (0.64 g, 2.10 mmol) of
bis-(p-nitrophenyl) carbonate in DMF (1 mL). This mixture
was stirred at room temperature for about 2 h, diluted with
ethyl acetate, extracted with HCl, NaOH, and brine. Yield
80%.
Compound 7. 150 mg of 6 in 3 mL of anhydrous dioxane was
combined with 0.35 mL of (S)-(+)-3-hydroxy-tetrahydrofuran
and 0.14 mL of triethylamine and refluxed for 48 h (yield
quantitative). ES+ 636.2 (M+1).
1H NMR (CDCl3) d 8.29 (2H, d), 7.91 (2H, d), 7.22 (5H, m),
5.13 (1H, m), 4.96 (1H, m), 4.52 (1H, d), 4.02 (1H, m), 3.84
(2H, m), 3.44 (1H, m), 3.36 (1H, m), 3.10 (3H, m, overlap),
2.88 (2H, m), 2.64 (1H, m), 2.14 (1H, m), 2.05 (1H, m), 1.84
(1H, m), 1.27 (9H, s), 0.78 (6H, two overl. d).
Compound 8: N-BuLi (1.0 mL, 2.5 mmol) and methyl chloro-
formate (0.0849 mL, 1.1 mmol) were added to 5 (0.521 g, 1
mmol) in tetrahydrofuran (5.0 mL) at 0 ꢁC under nitrogen
atmosphere. The mixture was gradually warmed to room
temperature over 3 h. After quenching with water and
extracting with ethyl acetate, the organic layer was dried over
MgSO4, filtered and concentrated under reduced pressure. The
residue was purified by silica gel chromatography with ethyl
acetate/hexane (1:2) to give 75 mg of the desired product 8
(13% yield).
1H NMR (400 MHz, CDCl3) d 8.35 (2H, d, J=8.5 Hz), 8.00
(2H, d, J=8.8 Hz), 7.37–7.21 (5H, m), 5.03 (1H, m), 4.61 (1H,
m), 4.07 (1H, m), 3.75 (3H, s), 3.55–3.34 (2H, m), 3.10 (1H,
m), 2.94 (2H, m), 2.71 (1H, m), 1.95 (1H, m), 1.30 (9H, s), 0.87
[6H, distorted doublet, J (apparent)=6.3 Hz]. LRMS: m/z
580.1 (MH+).
Compound 10: Dichloromethane (2 mL) and trifluoroacetic
acid (2 mL) were added to 8 (0.035 g, 0.0604 mmol), and the
mixture stirred at room temperature for 20 min.
1H NMR (400 MHz, CDCl3) d 8.34 (2H, d, J=8.6 Hz), 7.97
(2H, d, J=8.6 Hz), 7.30 (5H, m), 5.02 (1H, m), 3.80 (1H, m),
Compound 3. Obtained by treatment of 7 with 50% TFA/
1
DCM for 30 min and removal of solvents. H NMR (metha-
nol-d4): d 8.39 (2H, d, J=8.9 Hz), 8.23 (2H, m), 8.04 (2H, d,

2526
W. M. Kazmierski et al. / Bioorg. Med. Chem. Lett. 13 (2003) 2523–2526
J=8.9 Hz), 7.40–7.29 (5H, m), 5.14 (1H, m), 5.09 (1H, m),
3.84 (1H, m), 3.78 (1H, m), 3.72 (3H, m), 3.46 (1H, m), 3.35
(1H, m), 3.04 (1H, m), 2.93 (1H, m), 2.84 (2H, m), 2.16 (1H,
m), 1.90 (1H, m), 1.84 (1H, m), 0.74 (6H, m)
The conversion of 3 to 4 was performed under conditions
described for 10.
Compound 4. 1H NMR: d 8.38 (2H, d, J=8.8 Hz), 8.05 (2H, d,
J=8.7 Hz), 7.33–7.14 (5H, m), 7.12 (1H, m), 4.95 (1H, m),
3.70 (2H, m), 3.60 (1H, m), 3.51 (1H, m), 3.52 (1H, m), 3.39
(1H, m), 3.36 (1H, m), 3.14 (1H, m), 3.07 (1H, m), 2.94 (2H,
m), 2.47 (1H, m), 2.05 (1H, m), 1.97 (1H, m), 1.76 (1H, m),
0.84 (6H, m).