Chemoenzymatic synthesis of ketomethylene tripeptide isosteres

Article abstract of DOI:10.1016/j.tetasy.2004.11.092

A chemoenzymatic method is described for the synthesis of a desired ketomethylene tripeptide isostere. The key step is an enzymatic hydrolysis, which removes the C-terminal ester-protecting group under mild conditions without epimerizing the existing ster

Full text of DOI:10.1016/j.tetasy.2004.11.092

Tetrahedron:
Asymmetry
Tetrahedron: Asymmetry 16 (2005) 699–703
Chemoenzymatic synthesis of ketomethylene tripeptide isosteres
Junhua Tao,* Shanghui Hu, Qingping Tian, Naresh Nayyar and Srin Babu
Chemical Research and Development, PfizerGlobal R&D, 10578 Science Center Drive, San Diego, CA 92121, USA
Received 3 November 2004; accepted 22 November 2004
Available online 26 January 2005
Abstract—A chemoenzymatic method is described for the synthesis of a desired ketomethylene tripeptide isostere. The key step is an
enzymatic hydrolysis, which removes the C-terminal ester-protecting group under mild conditions without epimerizing the existing
stereogenic center and produces the desired stereoisomer in high enantiomeric excess (98% de, 97% ee). The method is short with
overall 15–20% yields after seven synthetic steps from readily available starting materials being obtained.
ꢀ 2005 Published by Elsevier Ltd.
1. Introduction
sired stereochemical configuration.1–10 Among them,
the shortest approach uses a chiral alkylation strategy
to produce both C- and N-terminal masked ketomethyl-
ene dipeptide isosteres from amino acids and (R)-2-hy-
droxy carboxylic esters.10 Assuming mild deprotection
conditions can be identified, these building blocks can
then be potentially incorporated into longer peptide se-
quences such as tripeptide isosteres (e.g., 1) and peptide
mimetics. For the synthesis of 1 and ultimately Rupintri-
virTM (Scheme 1),11 this methodology is faced with sev-
eral issues. First of all, the desired enantiomerically
pure (R)-3-(4-fluorophenyl)-2-hydroxypropionic ester 2
is not commercially available and its synthesis requires
Ketomethylene peptide isosteres, where a scissile amide
bond is replaced with a nonscissile carbon–carbon bond,
are recurring motifs used to design a variety of therapeu-
tic compounds with enhanced stability toward proteoly-
sis.1–9 For example, RupintrivirTM, which contains
ketomethylene tripeptide isostere 1 (Scheme 1), is a lead
candidate entering human clinical trials to treat the com-
mon cold infection.7
A number of methods have been reported for the prep-
aration of ketomethylene peptide isosteres with the de-
O
O
HO
O
H
CbzHN
C-5
OEt
OEt
F
O
F
F
3
2
4-fluorobenzaldehyde
50% yields from 2
40-45% for 4 steps
O
NH
O
O
O
O
N
O
N
H
OH
N
N
N
H
O
O
H
O
OEt
1
O
F
F
overall 10% yield
for 10 steps
Rupintrivir
Scheme 1.
*
Corresponding author. Tel.: +1 858 622 3228; fax: +1 858 678 8199; e-mail: junhua.tao@pfizer.com
0957-4166/$ - see front matter ꢀ 2005 Published by Elsevier Ltd.
doi:10.1016/j.tetasy.2004.11.092

700
J. Tao et al. / Tetrahedron: Asymmetry 16 (2005) 699–703
four steps with overall only 40–45% yields from 4-flu-
orobenzaldehyde.12 Secondly, the chiral alkylation step
usually gives poor yields (40–52%).10 After extensive
optimization, the best yield obtained for the dipeptide
isostere 3 is 55%. The major impurity results from com-
petitive O-alkylation.13 Thirdly, significant epimeriza-
tion at the existing stereogenic center C-5 occurred
under a variety of chemical conditions to remove the
ethyl protecting group in 3. The difficulty in removing
the minor wrong enantiomer makes the elongation from
the C-terminal problematic. Consequently, by using this
protocol, the synthesis of 1 requires ten steps with an
overall 10% only yield starting from 4-fluoro-
benzaldehyde.11
stituted alkene to generate the desired stereogenic center
at C-2 in 9 through either stereoinduction or chiral li-
gands. However, all of the tested conditions gave either
low yields or poor stereoselectivity.17 Finally standard
hydrogenation using Pd/C was applied to produce a mix-
ture of two diastereomers (2R,5S)- and (2S,5S)-9 in good
yields (95%). It should be noted that all attempts to re-
move the ester group in 9 by chemical hydrolysis led to
extensive epimerization at the stereogenic center C-5.
At this stage, enzymes were screened to catalytically re-
solve the diastereomeric mixture of 9 through ester
hydrolysis (Scheme 3). The identification of appropriate
enzymes relied on automated enzyme screening tech-
niques reported recently in our lab.18 A 96-well plate
kit, including lipases, proteases, acylases, esterases, and
amidases, was incubated with a diastereomeric mixture
of 9 on a thermomixer for 4–6 h. The reactions were
then quenched with acetonitrile and the results analyzed
by HPLC. A lipase from the fungus Mucor meihei was
identified as the best hit for a stereoselective hydrolysis
of 9 to produce the desired acid (2R,5S)-10. It is worth
noting that the enzyme is inexpensive thus allowing its
use for large-scale production. Substrate concentration
and pH were then examined for optimal conditions. It
was found that 10 could be produced in good resolution
yield (45%) and high enantiomeric excess (98.5% de,
97% ee, E = 170) at pH 7.5. Under these conditions,
no epimerization at stereogenic center C-5 was identified
in contrast to chemical hydrolysis. Finally, the target
compound 1 was obtained in good yields (70–75%)
and high enantiomeric excess (98% de, 97% ee) over
two steps via hydrogenolysis of 10 followed by acylation
with 11 under Schotten–Baumann conditions.19 The de-
sired stereochemistry at C-2 and C-5 was confirmed by
comparing to the literature data.7,8
These issues promoted us to look for alternative routes
and we describe herein a chemoenzymatic method for
the synthesis of tripeptide isostere 1. The method is
short with high yields and excellent enantiomeric
excesses.
2. Results and discussion
The synthesis starts with O-benzyl (S)-valine ester 5,
which can be readily prepared from (S)-valine 4 in good
yields (Scheme 2),14 followed by acyl transfer to the carb-
anion of a methyl phosphonate to give b-ketophospho-
nate 6. Subsequently, under Horner–Emmons
conditions, this compound was treated with NaH in
THF and reacted with a-ketoester 7, which was prepared
in one pot from 4-fluorobenzyl chloride in good yields
(85%).15,16 The resulting a,b-unsaturated c-ketoester 8
was obtained as a mixture of E- and Z-isomers (5:1) with
an overall 54% yield for three steps. Initial efforts were
directed toward stereocontrolled reduction of this trisub-
O
P
OMe
O
O
O
P
OMe
O
LiCH2
OMe
BnBr, K2CO3
OBn
OMe
OH
EtOH-H2O
N(Bn)2
N(Bn)2
NH2
5, 85%
4
6, 85%
F
F
O
O
1. NaH
F
10% Pd/C
H2
OEt
OEt
2.
O
N(Bn)2
O
N(Bn)2
O
OEt
E/Z = 5:1
8, 75%
(2R,5S/2S,5S)-9
O
95%
7
Scheme 2.
F
O
OH
Mucor miehei lipase
1. H2, Pd-C
1
9
N(Bn)2
O
O
2.
overall 15-20% yields
for 7 steps
N
(2R,5S)-10
45% yield
Cl
O
98.5% de, 97% ee
11
Scheme 3.

J. Tao et al. / Tetrahedron: Asymmetry 16 (2005) 699–703
701
By this approach, the desired ketomethylene tripeptide
isostere 1 was synthesized in 15–20% yields overall, over
seven steps starting from (S)-valine 4. The process has
shorter sequences and higher yields than the chiral alkyl-
ation route (15–20% vs 10%). The key step is an enzyme
catalyzed ester hydrolysis, which produces the desired
stereoisomer without epimerization of the existing stereo-
genic center at C-5.
then dispensed into each well via a multi-channel pip-
ette. ACN solution (10 lL) of the substrate (10–20 mg
of 9 per milliliter) was then added to each well, and
the plate incubated at 30 ꢁC on a thermomixer
(900 rpm). The reactions were quenched with 100 lL
of acetonitrile after 5 h. The 96-well plate was then cen-
trifuged, and the supernatant transferred from each well
into another 96-well plate and analyzed with automated
HPLC using a short C18 column (30 · 4.6 mm, 3 lm)
and ACN/H2O containing 0.1% TFA as the solvent
system.
3. Conclusion
In summary, a chemoenzymatic method has been devel-
oped for the synthesis of a ketomethylene tripeptide iso-
stere. This process has higher yields and a shorter
sequence than reported methods. This methodology
can potentially be extended to the preparation of a vari-
ety of ketomethylene peptide isosteres with different sub-
stituents at C-2 since a-ketoesters can be readily
prepared in one step with excellent yields (85–98%) from
alkyl or aryl halides.16 For example, branched substitu-
ents might be installed at C-2 with the desired stereo-
chemistry, where poor diastereoselectivity would be
expected using the chiral alkylation protocol.10
4.2. Tribenzyl ester 5
This compound was prepared according to a literature
protocol.14 1H NMR (300 MHz, CDCl3): d 7.32 (m,
15H), 5.29 (d, J = 12.3 Hz, 1H), 5.15 (d, J = 12.3 Hz,
1H), 3.97 (d, J = 14.1 Hz, 2H), 3.29 (d, J = 14.1 Hz,
2H), 2.15 (m, 1H), 1.01 (d, J = 6.3, 3H), 0.76 (d,
J = 6.3 Hz, 3H). 13C NMR (75 MHz, CDCl3): d 172.2,
140.0, 130.1, 129.0, 128.8, 128.0, 68.6, 65.6, 55.1, 26.7,
20.0, 16.4.
4.3. b-Ketophosphonate 6
To a solution of dimethyl methyl phosphonate (12.4 g,
100 mmol) in THF (80 mL) at À78 ꢁC was added drop-
wise BuLi (2.5 M in hexanes, 40 mL, 100 mmol). After
30 min, tribenzyl ester 5 (6.5 g, 16.7 mmol) in THF
(10 mL) was added and the mixture stirred for 20 min
at the same temperature followed by 3 h at rt. The reac-
tion was then quenched with satd NH4Cl (200 mL), and
extracted with MTBE (100 mL · 2). The combined
organic layer was washed with brine (60 mL), dried over
MgSO4, and concentrated to afford 6 as a crude product
(5.6 g, 13.9 mmol, 85% in yields, 90–95% UV purity),
which was carried over to the next step without further
4. Experimental
The majority of enzymes utilized in the preparation of
screening kits were obtained from various enzyme sup-
pliers including Amano (Nagoya, Japan), Roche (Basel,
Switzerland), Novo Nordisk (Bagsvaerd, Denmark),
Altus Biologics Inc. (Cambridge, MA), Biocatalytics
(Pasadena, CA), Toyobo (Osaka, Japan), Sigma–Aldrich
(St. Louis, MO, USA). HPLC analysis of the screened
samples was performed on an Agilent 220 HPLC auto-
sampler. Reactions were performed in an Eppendorf
thermomixer-R (VWR). Solvents used during optimiza-
tion were obtained from EM Science (Gibbstown, NJ)
and were of the highest purity available. Chiral HPLC
columns used in analysis were obtained from Chiral
Technologies (Exton, PA) and Phenomenex (Torrance,
CA). Each HPLC sample was prepared by taking
5 · 200 lL from the reaction slurry and then combined
and diluted with 4 mL of acetonitrile. 100 lL of that
solution was further diluted with 400 lL of acetonitrile
and injected into HPLC. For compound 1, the following
chiral HPLC method was used: Chiralcel OJ-R
(4.6100 mm, 3 lm); flow rate 0.5 mL/min; injection vol-
ume 10 lL; mobile phases: (A) 25 mM NaH2PO4
pH 2.0; (B) acetonitrile; isocratic: 40% B for 31 min,
detection at 254 nm. LC–MS conditions were as follows:
ESI (positive and negative modes); LC chromatography
runs on Luna C18 column (30 · 4.6 mm, 3 lm) using
ACN/H2O containing 0.1% formic acid as mobile
phases; gradient 10–90% ACN in 3.5 min, then isocratic
with 90% ACN for 1 min; flow rate 0.8 mL/min; detec-
tion at 254 nm.
1
purification. H NMR (300 MHz, CDCl3): d 7.28 (m,
10H), 2.80–4.2 (sets of m, 13H), 2.33 (m, 1H), 1.14 (d,
J = 6.0 Hz, 3H), 0.80 (d, J = 6.0 Hz, 3H). 13C NMR
(75 MHz, CDCl3): d 202.2, 139.5, 132.8, 129.0, 127.3,
71.1, 54.5, 52.8, 41.8, 40.1, 26.9, 19.0. ESI m/z [M+H]+
404.2.
4.4. b-Ketoester 7
This compound was prepared according to a literature
protocol with consistent NMR data.15,16 IR (KBr)
3421, 2974, 1725, 1504, 1221 cmÀ1. FABMS m/z
[M+H]+ 211.
4.5. c-Ketoester 9
To a stirred solution of NaH (60% in mineral oil, 0.7 g,
17.6 mmol) at 0 ꢁC was added slowly a solution of 6
(6.5 g, 16.0 mmol) in THF (80 mL). A solution of 7
(3.7 g, 17.6 mmol) in THF (15 mL) at the same temper-
ature was subsequently added. The ice bath was then re-
moved and the reaction kept at rt. After 18 h, the
reaction was quenched by water (50 mL) and extracted
with MTBE (50 mL · 2). The combined organic layer
was washed with brine (50 mL), dried over MgSO4,
and concentrated to afford 8 (5.5 g, 75%) as a crude
4.1. General procedure for enzyme screening
A screening kit (see Ref. 18) was thawed for 5 min and
80 lL of potassium phosphate buffer (0.1 M, pH 7.2)

702
J. Tao et al. / Tetrahedron: Asymmetry 16 (2005) 699–703
red oil. Without further purification, the crude 8 (4 g,
about 8.2 mmol) was dissolved in ethanol (40 mL), and
the mixture hydrogenated in the presence of 10% Pd/C
at 1 atm for 5 h. The reaction mixture was filtered
through Celite, which was washed with ethanol
(10 mL) and the combined organic layer concentrated
to give 9 (5.8 g, 12 mmol, ca. 75% over two steps from
6) upon chromatographic purification (hexanes/ethyl
acetate 95:5) over three steps as a mixture of two diaste-
reomers (1:1). 1H NMR (300 MHz, CDCl3): d 7.17–6.87
(m, 14H) 4.10–3.96 (m, 2H), 3.86 (dd, J = 6.0, 9.0 Hz,
2H), 3.37 (d, J = 14.4 Hz, 1H), 3.22 (d, J = 14.4 Hz,
1H), 3.05 (m, 1H), 2.89 (d, J = 10.8 Hz, 1H), 2.92 (d,
J = 10.5 Hz, 1H), 2.65–2.85 (m, 1H), 2.49–2.59 (m,
1H), 2.16–2.38 (m, 2H), 1.20 (m, 6H), 1.05 (m, 3H),
0.70 (d, J = 6.3 Hz, 3H), 0.60 (d, J = 6.3 Hz, 3H). 13C
NMR (75 ppm, CDCl3): d 210.93, 210.84, 174.94,
174.84, 140.11, 139.89, 134.76, 134.57, 130.89, 130.79,
129.09, 128.97, 128.73, 127.41, 115.93, 115.50, 70.88,
70.76, 61.06, 54.94, 54.79, 47.98, 47.45, 41.93, 41.64,
37.41, 37.36, 27.74, 27.64, 20.62, 20.39, 20.36, 14.41,
14.53.
To the solution was added slowly DIPEA (0.45 mL,
2.60 mmol) in dioxane (5 mL), followed by slow addi-
tion of an isoxazole acid chloride 11 (153 mg,
1.06 mmol) in dioxane (10 mL). After 2 h at the same
temperature, the reaction was finished as indicated by
HPLC and TLC (methylene chloride/hexanes/
IPA = 79/20/3) and to this mixture added methylene
chloride (20 mL). The organic layer was separated,
washed with 1 M HCl (10 mL), satd NaHCO3
(10 mL), dried over MgSO4, and filtered through a short
pad of silica gel to give 1 (216 mg, 0.56 mmol, about
72%, 98% de, 97% ee) upon chromatographic purifica-
tion (methylene chloride/hexane/IPA = 79/20/3). 1H
NMR (300 MHz, CDCl3): d 7.31 (d, J = 9.0 Hz, 1H),
7.14 (dd, J = 5.6, 8.5 Hz, 2H), 6.99 (t, J = 9.0 Hz, 2H),
6.40 (s, 1H), 4.70 (dd, J = 4.3, 8.6 Hz, 1H), 3.20 (m,
1H), 3.07 (dd, J = 6.0, 13.7 Hz, 1H), 2.98 (t,
J = 9.4 Hz, 1H), 2.76 (dd, J = 8.3, 13.6 Hz, 1H), 2.56
(dd, J = 4.4, 18.3 Hz, 1H), 2.48 (s, 3H), 2.32 (m, 1H),
1.02 (d, J = 6.8 Hz, 3H), 0.85 (d, J = 6.8 Hz, 3H). 13C
NMR (75 MHz, CDCl3): d 206.28, 179.19, 171.73,
159.83, 158.57, 133.98, 130.86, 116.01, 115.73, 101.75,
62.77, 41.77, 36.91, 30.64, 20.27, 17.04, 12.61. ESI
[MÀH]À m/z: 389.2.
4.6. (2R,5S)-Acid 10
To a 150 mL jacketed flask equipped with a pH elec-
trode, ester 9 (3 g, 6.13 mmol) in heptane (12 mL) was
added followed by Mucor miehei lipase (30 mL, Sigma
brand) in 18 mL of phosphate buffer (pH 7.5, 0.5 M).
The reaction mixture was stirred at 25 ꢁC at pH 7.5,
which was maintained by an auto pH titrator. The con-
version and diastereoselectivity was monitored by
HPLC. When the conversion approached 45% (ca.
48 h), the reaction mixture was adjusted to pH 4.5 and
extracted with heptane (60 mL · 2). The heptane layer
was combined and extracted with methanol (containing
0.5% water) and the methanolic phase washed with hep-
tane. After removal of methanol, the residue was slur-
ried in hexane to afford the desired acid 10 (1.3 g,
2.81 mmol, 45% in yield, de 98.5%, ee 97%, 96% purity).
1H NMR (300 MHz, CD3OD): d 7.05–7.38 (m, 14H),
3.36 (s, 2H), 3.27 (d, J = 13.8 Hz, 1H), 3.05 (m, 2H),
2.92 (d, J = 10.5 Hz, 1H), 2.54–2.66 (m, 2H), 2.40 (dd,
J = 3.3, 18.9 Hz, 1H), 2.15 (m, 1H), 1.19 (m, 1H), 1.05
(d, J = 6.6 Hz, 3H), 0.74 (d, J = 6.3 Hz, 3H). 13C
NMR (75 ppm, CD3OD): d 210.78, 178.50, 139.81,
135.75, 130.80, 128.81, 128.28, 127.05, 115.26, 114.98,
70.08, 63.81, 63.16, 54.53, 42.85, 37.16, 27.34, 19.80,
19.44. ESI [M+H]+ m/z: 462.2.
Acknowledgements
The authors thank Kim Albizati, Van Martin and
Chunrong Ma for general discussions, Jason Ewanicki
for NMR and Xiaobing Xiong for HPLC support.
References
1. Neffe, A. T.; Meyer, B. Angew. Chem., Int. Ed. 2004, 43,
2937–2940.
2. Hansen, K. K.; Grosch, B.; Greiveldinger-Poenaru, S.;
Bartlett, P. A. J. Org. Chem. 2003, 68, 8465–8470.
3. Ebetino, F. H.; Mazur, A. W.; Hayes, J. C.; Wang, F.;
Solinsky, M. G.; Colson, A.-O.; Lin, Q. PCT Int. Appl.
2002, 123, WO 2002026774.
4. Shearman, M. S.; Beher, D.; Clarke, E. E.; Lewis, H. D.;
Harrison, T.; Hunt, P.; Nadin, A.; Smith, A. L.; Steven-
son, G.; Castro, J. L. Biochemistry 2000, 39, 8698–
8704.
5. Blakskjr, P.; Hoj, B.; Riber, D.; Skrydstrup, T. J. Am.
Chem. Soc. 2003, 125, 4030–4031.
6. Dias, L. C.; Diaz, G.; Ferreira, A. A.; Meira, P. R.;
Ferreira, E. Synthesis 2003, 4, 603–622.
7. Webber, S. E.; Okano, K.; Little, T. L.; Reich, S. H.; Xin,
Y.; Fuhrman, S. A.; Matthews, D. A.; Love, R. A.;
Hendrickson, T. F.; Patick, A. K.; Meador, J. W., III;
Ferre, R. A.; Brown, E. L.; Ford, C. E.; Binford, S. L.;
Worland, S. T. J. Med. Chem. 1998, 41, 2786.
8. Tao, J.; Babu, S.; Dagnino, R.; Tian, R.; Remarchuk, T.;
McGee, K.; Nayyar, N.; Moran, T. PCT Int. Appl. 2001,
127.
9. Tian, Q.; Nayyar, N. K.; Babu, S.; Chen, L.; Tao, J.; Lee,
S.; Tibbetts, A.; Moran, T.; Liou, J.; Guo, M.; Kennedy,
T. P. Tetrahedron Lett. 2001, 42, 6807.
10. Hoffman, R. V.; Tao, J. Tetrahedron 1997, 53, 7119, and
references cited therein; For earlier synthesis of nonke-
4.7. Tripeptide isostere 1
Compound 10 (680 mg, about 1.06 mmol) was dissolved
into EtOH (10 mL) followed by H2SO4 (116 mg,
1.10 mmol) and 10% Pd–C (204 mg) in a Parc-shaker
(50 psi). After about 5 h, the reaction finished as indi-
cated by HPLC and TLC, and the mixture was then dis-
solved into methanol (75 mL), filtered through a short
pad of Celite, which was then washed with another ali-
quot of methanol (75 mL). The combined solvent was
evaporated to give a crude colorless oil (490 mg), with-
out further purification, which was then dissolved into
dioxane (40 mL) and the mixture then cooled to 10 ꢁC.
´
tomethylene carboxyl modified peptides, see: Sauve, G.;
Berre, N. L.; Zacharie, B. J. Org. Chem. 1990, 55, 3002–
3004.
11. Tao, J., unpublished.

J. Tao et al. / Tetrahedron: Asymmetry 16 (2005) 699–703
703
12. For a detailed description of the synthesis of compound 2,
see: Tao, J.; McGee, K. Org. Process Res. Dev. 2002, 6, 520.
13. The best ratio of C- over O-alkylation was found to be 3:1.
The reaction is highly dependent upon solvents and bases.
For example, using NaH, the reaction is complete within
16–24 in THF (3:1 C-/O-alkylation), while the use of t-
BuOK or Cs2CO3 leads to the consumption of starting
materials within 1–5 h (2:1 C-/O-alkylation). The alkyl-
ation takes place faster in more polar solvents than less
polar ones. For example, no products were detected in
toluene and methylene chloride at rt after 24 h.11
15. Schuster, H.; Coppola, G. J. Heterocycl. Chem. 1994, 31,
1381.
16. Singh, J.; Kissick, T. P.; Mueller, R. H. Org. Prep. Proced.
Int. 1989, 21, 501–504.
17. For example, poor activities were identified for CrabtreeÕs,
NoyoriÕs ligands, WilkinsonÕs catalyst, catecholborane,
catecholborane/(PPh3)3RhCl, LiAlH4/CuI, Zn/AcOH, and
Me2CuLi. Sodium borohydride mostly gave the product
resulting from the ketone reduction, while StrykerÕs
reagent led to poor stereoselectivity.11
18. Yazbeck, D. R.; Tao, J.; Martinez, C. A.; Kline, B. J.; Hu,
S. Adv. Synth. Catal. 2003, 345, 524.
19. Charvillon, F.; Amouroux, R. Tetrahedron Lett. 1996, 37,
5103.
14. Bonjoch, J.; Catena, J.; Terricabras, D.; Fernandez, J.-C.;
Lopez-Canet, M.; Valls, N. Tetrahedron: Asymmetry 1997,
8, 3143–3151.