Incorporation of trifluoroisoleucine into proteins in vivo

Article abstract of DOI:10.1021/ja0298287

Two fluorinated derivatives of isoleucine: D,L-2-amino-3-trifluoromethyl pentanoic acid (3TFI, 2) and D,L-2-amino-5,5,5-trifluoro-3-methyl pentanoic acid (5TFI, 3) were prepared. 5TFI was incorporated into a model target protein, murine dihydrofolate reductase (mDHFR), in an isoleucine auxotrophic Escherichia coli host strain suspended in 5TFI-supplemented minimal medium depleted of isoleucine. Incorporation of 5TFI was confirmed by tryptic peptide analysis and matrix-assisted laser desorption ionization mass spectrometry (MALDI-MS) of the protein product. Amino acid analysis showed that more than 93% of the encoded isoleucine residues were replaced by 5TFI. Measurement of the rate of activation of 5TFI by the E. coli isoleucyl-tRNA synthetase (IleRS) yielded a specificity constant (k_cat/K_m) 134-fold lower than that for isoleucine. 5TFI was successfully introduced into the cytokine murine interleukin-2 (mIL-2) at the encoded isoleucine positions. The concentration of fluorinated protein that elicits 50% of the maximal proliferative response is 3.87 ng/mL, about 30% higher than that of wild-type mIL-2 (EC₅₀ = 2.70 ng/mL). The maximal responses are equivalent for the fluorinated and wild-type cytokines, indicating that fluorinated proteins can fold into stable and functional structures. 3TFI yielded no evidence for in vivo incorporation into recombinant proteins, and no evidence for activation by IleRS in vitro.

Full text of DOI:10.1021/ja0298287

Published on Web 05/17/2003
Incorporation of Trifluoroisoleucine into Proteins in Vivo
Pin Wang, Yi Tang, and David A. Tirrell*
Contribution from the DiVision of Chemistry and Chemical Engineering, California Institute of
Technology, Pasadena, California 91125
Received December 19, 2002; E-mail: tirrell@caltech.edu
Abstract: Two fluorinated derivatives of isoleucine: D,L-2-amino-3-trifluoromethyl pentanoic acid (3TFI, 2)
and D,L-2-amino-5,5,5-trifluoro-3-methyl pentanoic acid (5TFI, 3) were prepared. 5TFI was incorporated
into a model target protein, murine dihydrofolate reductase (mDHFR), in an isoleucine auxotrophic
Escherichia coli host strain suspended in 5TFI-supplemented minimal medium depleted of isoleucine.
Incorporation of 5TFI was confirmed by tryptic peptide analysis and matrix-assisted laser desorption ionization
mass spectrometry (MALDI-MS) of the protein product. Amino acid analysis showed that more than 93%
of the encoded isoleucine residues were replaced by 5TFI. Measurement of the rate of activation of 5TFI
m
by the E. coli isoleucyl-tRNA synthetase (IleRS) yielded a specificity constant (kcat/K ) 134-fold lower than
that for isoleucine. 5TFI was successfully introduced into the cytokine murine interleukin-2 (mIL-2) at the
encoded isoleucine positions. The concentration of fluorinated protein that elicits 50% of the maximal
proliferative response is 3.87 ng/mL, about 30% higher than that of wild-type mIL-2 (EC50 ) 2.70 ng/mL).
The maximal responses are equivalent for the fluorinated and wild-type cytokines, indicating that fluorinated
proteins can fold into stable and functional structures. 3TFI yielded no evidence for in vivo incorporation
into recombinant proteins, and no evidence for activation by IleRS in vitro.
Introduction
Protein engineering provides powerful tools for modification
Multisite replacement is complementary to the site-specific
method of analogue insertion, and is of primary interest in
situations where one wishes to change the overall physical
1,2
of natural proteins and for de novo design of artificial proteins.
behavior of the engineered protein.
Among the most recently developed protein engineering tools
Fluorinated amino acids are of special interest.18 Fluorine is
not a sterically demanding subsituent (van der Waals radii
r(F) ) 1.35 Å, r(H) ) 1.20 Å) and although the C-F bond is
2
are methods for incorporation of noncanonical amino acids.
Novel amino acids can be introduced by chemical methods such
3
as solid-phase peptide synthesis and native chemical ligation.
0
.4 Å longer than the C-H bond, introduction of fluorine
Expressed protein ligation and in vitro translation methods allow
1
9
generally causes minimal structural perturbation. Fluorocar-
bons yield higher contact angles with water compared to their
hydrocarbon forms, indicating their elevated hydrophobicity.20
Fluorination is commonly employed in drug delivery to generate
4
-6
site-specific incorporation of nonnatural amino acids.
Most
recently, there has emerged an apparently general strategy for
site-specific incorporation of noncanonical amino acids into
proteins in vivo.7
-9
21
22
stable liposomes and in drug design to increase drug activity,
We have previously exploited the ability of auxotrophic
Escherichia coli host strains to introduce amino acid analogues
into proteins in multisite fashion. Functional groups distinct from
those of the natural amino acids, including alkenes, alkynes,
aryl halides, azides, ketones, pyridyls, and others have been
both of which result at least in part from the enhanced
hydrophobic character of fluorinated substituents. In light of
fact that hydrophobic forces play important roles in protein
(
12) Sharma, N.; Furter, R.; Kast, P.; Tirrell, D. A. FEBS Lett. 2000, 467, 37-
1
0-17
incorporated into proteins in vivo with good efficiency.
40.
(13) Kirshenbaum, K.; Carrico, I. S.; Tirrell, D. A. Chembiochem 2002, 3, 235-
237.
(
1) Dolgikh, D. A.; Kirpichnikov, M. P.; Ptitsyn, O. B.; Chemeris, V. V. Mol.
(14) Kothakota, S.; Mason, T. L.; Tirrell, D. A.; Fournier, M. J. J. Am. Chem.
Soc. 1995, 117, 536-537.
Biol. 1996, 30, 149-156.
(
(
(
2) van Hest, J. C. M.; Tirrell, D. A. Chem. Commun. 2001, 1897-1904.
3) Dawson, P. E.; Kent, S. B. H. Annu. ReV. Biochem. 2000, 69, 923-960.
4) Bain, J. D.; Glabe, C. G.; Dix, T. A.; Chamberlin, A. R.; Diala, E. S. J.
Am. Chem. Soc. 1989, 111, 8013-8014.
(15) Kiick, K. L.; van Hest, J. C. M.; Tirrell, D. A. Angew. Chem., Int. Ed.
Engl. 2000, 39, 2148+.
(16) Kiick, K. L.; Saxon, E.; Tirrell, D. A.; Bertozzi, C. R. Proc. Natl. Acad.
Sci. USA 2002, 99, 19-24.
(
5) Noren, C. J.; Anthonycahill, S. J.; Griffith, M. C.; Schultz, P. G. Science
(17) Datta, D.; Wang, P.; Carrico, I. S.; Mayo, S. L.; Tirrell, D. A. J. Am. Chem.
Soc. 2002, 124, 5652-5653.
1
989, 244, 182-188.
(
(
(
(
6) Hecht, S. M. Acc. Chem. Res. 1992, 25, 545-555.
(18) Kukhar, V. P.; Soloshonok, V. A. Fluorine Containing Amino Acids:
Synthesis and Properties; John Wiley & Sons.: Chichester, 1995.
(19) Resnati, G. Tetrahedron 1993, 49, 9385-9445.
(20) Sheppard, W. A.; Sheetz, C. M. Organic Fluorine Chemistry; W. A.
Benjamin, Inc.: New York, 1969.
7) Furter, R. Protein Sci. 1998, 7, 419-426.
8) Wang, L.; Schultz, P. G. Chem. Commun. 2002, 1-11.
9) Wang, L.; Brock, A.; Schultz, P. G. J. Am. Chem. Soc. 2002, 124, 1836-
1
837.
(
(
10) van Hest, J. C. M.; Tirrell, D. A. FEBS Lett. 1998, 428, 68-70.
11) van Hest, J. C. M.; Kiick, K. L.; Tirrell, D. A. J. Am. Chem. Soc. 2000,
(21) Krafft, M. P.; Riess, J. G. Biochimie 1998, 80, 489-514.
(22) Takagi, Y.; Nakai, K.; Tsuchiya, T.; Takeuchi, T. J. Med. Chem. 1996,
39, 1582-1588.
1
22, 1282-1288.
6900
9
J. AM. CHEM. SOC. 2003, 125, 6900-6906
10.1021/ja0298287 CCC: $25.00 © 2003 American Chemical Society

Trifluoroisoleucine into Proteins in Vivo
A R T I C L E S
folding and protein-protein recognition, it was suggested in a
recent review that fluorinated amino acids could be utilized in
protein engineering to prepare peptides and proteins with
enhanced structural stability.23 We and Kumar and co-workers
have recently demonstrated that substitution of leucine residues
by 5,5,5-trifluoroleucine at the d-positions of the leucine zipper
Scheme 1
peptide GCN4-p1 increases the thermal stability of the coiled-
coil structure.2
4-27
Furthermore, we have been able to show that
the fluorinated peptides maintain affinity and specificity in
binding to their target DNA sequences.25 To date, it has been
2
8
shown that mono-fluorinated analogues of phenylalanine,
29
30
tryptophan, and tyrosine, as well as trifluoroleucine, trifluo-
romethionine, and hexafluoroleucine32 are translationally ac-
tive. Each of these amino acids can be readily incorporated into
target proteins in appropriate auxotrophic E. coli strains, though
incorporation of hexafluoroleucine requires overexpression of
the leucyl-tRNA synthetase of the host.32 In this work, we
prepared two fluorinated analogues of isoleucine (1): D,L-2-
amino-3-trifluoromethyl pentanoic acid (2, 3TFI) and D,L-2-
amino-5,5,5-trifluoro-3-methyl pentanoic acid (3, 5TFI). We find
that 3, but not 2, serves as an excellent surrogate for isoleucine
in supporting protein synthesis in E. coli. Incorporation of 3
into the five isoleucine sites of murine interleukin-2 (mIL-2)
yields a fluorinated variant characterized by activity nearly
identical to that of the authentic cytokine.
31
Scheme 2
was removed in vacuo. The light yellow filtrate was subjected to
vacuum distillation to afford 10 g (51 mmol, 63%) of a mixture of the
E and Z forms of 3-trifluoromethyl-pent-2-enoic acid ethyl ester (94:6
by NMR). NMR assignments for the E and Z isomers were made
Experimental Section
according to previous reports on the synthesis of 4,4,4-trifluoro-3-
methyl-but-2-enoic acid ethyl ester.4
1,42 1
H NMR (CDCl
-C(CF )dCH), 2.67 (q,
)dCH), 4.2 (q, 2H, CH -CH -O), 6.23 (s, 1H,
)dCH); (Z) δ 1.06 (t, 3H, CH -CH -O), 1.23 (t,
-C(CF )dCH), 2.32 (q, 2H, CH -CH -C(CF )dCH),
.2 (q, 2H, CH -CH -O), 6.05 (s, 1H, CH -CH -C(CF )dCH).
3-Trifluoromethylpentanoic Acid Ethyl Ester. 3-Trifluorometh-
ylpent-2-enoic acid ethyl ester (50 g, 255 mmol) was dissolved in 100
mL of ethanol. PtO (1.5 g) was added, and the mixture was
hydrogenated at 10 psi at room temperature for 12 h. The catalyst was
removed by filtration, 1 g of new PtO was added and hydrogenation
): (E) δ 1.06
3
Synthesis of Amino Acid Analogues. D,L-2-Amino-5,5,5-trifluoro-
-methyl pentanoic acid (3) was prepared according to Muller et al.33
(t, 3H, CH
H, CH -CH
CH -CH -C(CF
H, CH -CH
3
-CH
2
-O), 1.23 (t, 3H, CH
3
-CH
2
3
3
2
3
2
-C(CF
3
3
2
This section provides detailed protocols for preparation of 2 (Scheme
3
2
3
3
2
1
) and NMR data for intermediates prepared en route to 3 (Scheme 2).
Materials and General Methods. Glassware was dried at 150 °C
3
4
3
2
3
3
2
3
3
2
3
2
3
and cooled under nitrogen prior to use. 1,1,1-trifluoro-butan-2-one (4)
was purchased from FluoroChem Inc. Unless otherwise specified, all
reagents were used as received. Hydrogenation was performed with a
Parr high-pressure hydrogenation apparatus (Westinghouse Electric
3
4
2
1
19
Corp.). H NMR and F NMR spectra were recorded on a Varian
Mercury 300 MHz NMR spectrometer.
2
E,Z-3-Trifluoromethylpent-2-enoic Acid Ethyl Ester. Carboeth-
oxymethylenetriphenyl-phosphorane (27.62 g, 80.5 mmmol), 1,1,1-
trifluoro-butan-2-one (4) (10 g, 79.4 mmol) and 70 mL ether were
placed in a dry, 250 mL three necked flask. Reaction occurred
immediately with formation of a thick paste. An additional 40 mL of
ether was added and the mixture was refluxed for 24 h at 50-60 °C
with mechanical stirring. The reaction mixture was filtered and ether
was continued for another 10 h. The catalyst was filtered out and ethanol
was removed in vacuo. Vacuum distillation yielded 35 g (176 mmol,
70%) of 3-trifluoromethylpentanoic acid ethyl ester. H NMR
1
3 3 2 3 2 3
(CDCl ): δ 1.02 (t, 3H, CH -CH -CH(CF )-CH ), 1.28 (t, 3H, CH -
(34) McBee, E. T.; Hsu, C. G.; Roberts, C. W. J. Am. Chem. Soc. 1956, 78,
3
389-3392.
(
35) Yamamoto, Y.; Shimokawa, K.; Yoshizawa, T.; Kawano, T.; Iwai, H.
Preparation and testing of 1-carbamoyl-5-fluorouracil deriVatiVes as
neoplasm inhibitors. Eur. Pat. Appl. 1988, p 13.
(
23) Marsh, E. N. G. Chem. Biol. 2000, 7, R153-R157.
(
24) Bilgicer, B.; Fichera, A.; Kumar, K. J. Am. Chem. Soc. 2001, 123, 4393-
(36) Marson, C. M.; Grabowska, U.; Fallah, A.; Walsgrove, T.; Eggleston, D.
S.; Baures, P. W. J. Org. Chem. 1994, 59, 291-296.
(37) Miller, J. H. A Short Course in Bacterial Genetics: A Laboratory Manual
and Handbook for Escherichia coli and Related Bacteria; Cold Spring
Harbor Laboratory: New York, 1992.
4
399.
(
25) Tang, Y.; Ghirlanda, G.; Vaidehi, N.; Kua, J.; Mainz, D. T.; Goddard, W.
A.; DeGrado, W. F.; Tirrell, D. A. Biochemistry 2001, 40, 2790-2796.
26) Bilgicer, B.; Xing, X.; Kumar, K. J. Am. Chem. Soc. 2001, 123, 11 815-
(
1
1 816.
(38) Hendrickson, T. L.; Nomanbhoy, T. K.; Schimmel, P. Biochemistry 2000,
39, 8180-8186.
(
27) Yoder, N. C.; Kumar, K. Chem. Soc. ReV. 2002, 31, 335-341.
(
28) Yoshikawa, E.; Fournier, M. J.; Mason, T. L.; Tirrell, D. A. Macromolecules
(39) Guisez, Y.; Demolder, J.; Mertens, N.; Raeymaekers, A.; Plaetinck, G.;
Robbens, J.; Vandekerckhove, J.; Remaut, E.; Fiers, W. Protein Expres.
Purif. 1993, 4, 240-246.
1
994, 27, 5471-5475.
(
(
29) Gerig, J. T. Prog. NMR Spectrosc. 1994, 26, 293-370.
30) Tang, Y.; Ghirlanda, G.; Petka, W. A.; Nakajima, T.; DeGrado, W. F.;
Tirrell, D. A. Angew. Chem., Int. Ed. Engl. 2001, 40, 1494+.
(40) Gillis, S.; Ferm, M. M.; Ou, W.; Smith, K. A. J. Immunol. 1978, 120,
2027-2032.
(
31) Duewel, H.; Daub, E.; Robinson, V.; Honek, J. F. Biochemistry 1997, 36,
(41) Poulter, C. D.; Satterwhite, D. M.; Rilling, H. C. J. Am. Chem. Soc. 1976,
98, 3376-3377.
3
404-3416.
(
32) Tang, Y.; Tirrell, D. A. J. Am. Chem. Soc. 2001, 123, 11 089-11 090.
(42) Dull, D. L.; Baxter, I.; Mosher, H. S. J. Org. Chem. 1967, 32, 1622-
1623.
(33) Muller, N. J. Fluorine Chem. 1987, 36, 163-170.
J. AM. CHEM. SOC.
9
VOL. 125, NO. 23, 2003 6901

A R T I C L E S
Wang et al.
CH
H, CH
CH(CF )-CH
q, 2H, CH -CH
-Trifluoromethylpentanoic Acid (5).35 3-Trifluoromethylpentanoic
acid ethyl ester (5 g, 22 mmol) was added to 50 mL of 1N of aqueous
NaOH solution and the mixture was refluxed for 24 h and then extracted
with 50 mL of ether. The solution was acidified with HCl and extracted
three times with 200 mL portion of ether. The combined ether layers
were dried and ether was removed in vacuo. Vacuum distillation
2
-O), 1.4-1.6 (m, 1H, CH
-CH -CH(CF )-CH
), 2.52-2.72 (m, 2H, CH
-O).
3
-CH
), 2.35-2.45 (m, 1H, CH
-CH -CH(CF )-CH
2
-CH(CF
3
)-CH
2
), 1.7-1.8 (m,
-CH
), 4.19
(d, allo-isoleucine form), δ 9.6 (d, isoleucine form). MS(EI) molecular
1
3
2
3
2
3
2
-
6 10 3 2
ion calcd for C H F NO 185.14, found 185.1.
36
3
2
3
2
3
2
3-Methyl-but-3-enenitrile (8). Cyanoacetic acid (100 g, 1.23 mol),
71.3 g (1.23 mol) of acetone, and 8.0 g (0.1 mol) of ammonium acetate
were dissolved in 200 mL of dry benzene and the solution was subjected
to reflux in a Dean-Stark apparatus overnight. The Dean-Stark
(
3
2
3
apparatus was replaced with a distillation head and the desired product
1
(
52 g, 55%) was collected between 110 and 115 °C. H NMR (CDCl
3
)
δ 1.7-2.0 (s, 3H, CH
C(CH )-CH -CN), 4.94 (s, H, CH
CH dC(CH )-CH -CN).
-Methyl-5,5,5-trifluoropentanonitrile (9).33 1H NMR (CDCl
.02/1.19 (d, 3H, CF -CH -CH(CH )), 2.0-2.3 (m, 3H, CF -CH
CH(CH )), 2.4 (d, 2H, CH(CH )-CH -CN).
2-Bromo-3-methyl-5,5,5-trifluoropentanoic acid.33 1H NMR (CDCl
δ 1.21 (dd, 3H, CF -CH -CH(CH )), 2.0-2.3 (m, 1H, CF -CH
)), 2.3-2.8 (m, 2H, CF -CH -CH(CH )), 4.21/4.5 (d, 1H,
)-CH(Br)), 11.6 (s, 1H, COOH).
2
dC(CH
3
2
)-CH -CN), 3.0 (s, 2H, CH
2
d
3
2
2
dC(CH
3
)-CH -CN), 5.06 (s, H,
2
1
2
3
2
afforded pure 5 (2.55 g, 15 mmol, 70%). H NMR (CDCl
H, CH -CH -CH(CF )-CH ), 1.54 (m, 1H, CH -CH -CH(CF
CH ), 1.76 (m, 1H, CH -CH -CH(CF )-CH ), 2.45 (m, 1H, CH
CH -CH(CF )-CH ), 2.65 (m, 2H, CH -CH -CH(CF )-CH ), 11.54
s, 1H, COOH).
Ethyl 2-bromo-3-trifluoromethylpentanoate. To 5 g (29.4 mmol)
3
) δ 1.02 (t,
3
3
) δ
3
3
2
3
2
3
2
3
)-
1
3
2
3
3
2
-
2
3
2
3
2
3
-
3
3
2
2
3
2
3
2
3
2
(
3
)
3
2
3
3
2
-
CH(CH
3
3
2
3
of 3-trifluoromethylpentanoic acid at 30 °C, 4.2 g (35.3 mmol) of
thionyl chloride was added dropwise. The mixture was stirred at 60-
CH(CH
3
D,L-2-Amino-5,5,5-trifluoro-3-methyl Pentanoic Acid (3). 1H
3
3
8
0 °C until gas evolution essentially ceased. At 80 °C, 5.64 g (35.3
NMR (D
CF -CH
d, 1H, CH(CH
1.54 (t, 3F, CF
found 185.1.
2
O) δ 0.96 (dd, 3H, CF
-CH(CH )), 2.25-2.5 (m, 2H, CF
)-CH(NH
)). 19F NMR (D
). MS(EI) molecular ion calcd for C
3
-CH
2
-CH(CH
-CH
O, TFA as reference) δ
NO 185.14,
3
)), 2.10-2.25 (m, 1H,
mmol) of Br was added dropwise at a rate that matched the rate of
2
3
2
3
3
2
-CH(CH )), 3.65
3
consumption. Stirring was continued for several hours until the evolution
of HBr stopped. The reaction mixture was cooled to room temperature
and 60 mL of absolute ethanol was added slowly to the crude acid
chloride. After standing overnight, the mixture was washed with water,
(
1
3
2
2
3
H F
6 10 3
2
Determination of Translational Activity. Buffers and media were
prepared according to standard protocols. An isoleucine auxotrophic
and the organic layer was dried over Na
2
SO
) δ 0.93 (t, 3H, CH
-CH -O), 1.48 (m, 1H, CH
CH(Br)), 1.72 (m, 1H, CH -CH -CH(CF )-CH(Br)), 2.35 (m, 1H,
CH -CH -CH(CF )-CH(Br)), 2.62 (m, 1H, CH -CH -CH(CF )-
CH(Br)), 4.18 (q, 2H, CH -CH -O).
4
. Yield 6.5 g (23 mmol,
-CH -CH(CF )-CH-
-CH -CH(CF )-
1
8
0%). H NMR (CDCl
3
3
2
3
-
derivative of E. coli strain BL21(DE3), designated AI (E. coli B F
(
Br)), 1.25 (t, 3H, CH
3
2
3
2
3
-
-
B
m ) gal dcm λ(DE3) ilVD691), constructed in our
ompT hsdS(r
laboratory, was used as host strain. Host cells of E. coli strain BW
159 containing the ilVD691:Tn10 mutation were first infected by
B
3
2
3
37
3
2
3
3
2
3
6
3
2
transducing phage P1 and the transposon was transferred into the
expression strain BL21(DE3) (Novagen). The resulting auxotroph was
treated with chlorotetracycline and fusaric acid to afford the stable
auxotroph AI. The repressor plasmid pLysS-IQ was constructed by
Ethyl 2-azido-3-trifluoromethylpentanoate (6). Ethyl 2-bromo-3-
trifluoromethylpentanoate (5 g, 18 mmol), 25 g (77 mmol) of sodium
azide, 10 mL of ethanol and enough water to dissolve the sodium azide
were refluxed for 6 days. The reaction mixture was distilled (CAU-
TION! Organic azides should be distilled with special care), and the
q
12
Sharma, and carried the lacI gene for lac repressor. The AI strain
with the repressor plasmid pLysS-IQ was designated AI-IQ. The
expression system AI-IQ[pQE15] was obtained by transformation of
pQE15 (Qiagen) into AI-IQ.
organic layer was used for the next step without further purification.
1
Yield 2.58 g (10.8 mmol, 60%). H NMR (CDCl
3
) δ 0.98 (t, 3H, CH
-CH -O), 1.48 (m, 1H,
)), 1.7 (m, 1H, CH -CH -CH(CF )-
-CH -CH(CF )-CH(N )), 2.5-2.7 (m,
)-CH(N )), 4.16 (q, 3H, CH -CH -O). IR
) 2100 cm (strong).
Ethyl 2-amino-3-trifluoromethyl-pentanoate Hydrochloride. Eth-
3
-
CH
2
-CH(CF
-CH -CH(CF
)), 2.36 (m, 1H, CH
H, CH -CH -CH(CF
CDCl
3
)-CH(N
3
)), 1.25 (t, 3H, CH
3
2
Protein Expression and Incorporation of Trifluoroisoleucine into
mDHFR. Small scale (5 mL) cultures were used to investigate the in
vivo translational activity of 2 and 3. M9 minimal medium (50 mL)
CH
3
2
3
)-CH(N
3
3
2
3
CH(N
3
3
2
3
3
1
(
3
2
3
3
3
2
4
supplemented with 0.2% glucose, 1 mg/L thiamine, 1 mM MgSO ,
-
1
3
0
.1 mM CaCl
2
, 19 amino acids (at 20 mg/L), antibiotics (ampicillin
2
00 mg/L, chloramphenicol 35 mg/L) and isoleucine (at 20 mg/L) was
yl 2-azido-3-trifluoromethylpentanoate (4 g, 16.7 mmol) was dissolved
in ethanol (8 mL). Pd/C (0.04 g, 10% Pd) was added as catalyst, and
the mixture was hydrogenated at low pressure (10 psi) over 12 h. The
catalyst was filtered out and HCl in ether (40 mL) was added, after
inoculated with 1 mL overnight culture of AI-IQ[pQE15]. When the
optical density at 600 nm reached 0.8-1.0, a medium shift was
performed. Cells were sedimented by centrifugation for 15 min at 4000
g at 4 °C, the supernatant was removed and the cell pellets were washed
twice with 0.9% NaCl. Cells were resuspended in supplemented M9
medium containing either: (a) 500 mg/L analogue (2 or 3), (b) 20 mg/L
Ile (1) (positive control), and (c) no Ile or analogues (negative control).
Protein expression was induced 10 min after the medium shift by
addition of isopropyl-â-D-thiogalactoside (IPTG) to a final concentration
of 1 mM. Cells were cultured for 4 h post-induction and protein
expression was monitored by SDS polyacrylamide gel electrophoresis
(PAGE, 12%), using a normalized OD600 of 0.2 per sample.
Larger scale expression of mDHFR was preformed in 1 L cultures
of AI-IQ[pQE15] grown in the supplemented M9 medium described
above. Cells were sedimented and washed twice with 0.9% NaCl and
resuspended in supplemented M9 medium containing 3 at 500 mg/L.
Gene expression was induced with 1 mM IPTG 10 min after the
medium shift.
which the solvent was removed in vacuo to afford crude product (1 g,
1
4
.7 mmol, 28%). H NMR (CDCl
3
) δ 1.02 (t, 3H, CH
-CH -O), 1.6-1.85 (m, 2H, CH
)), 2.97 (m, 1H, CH -CH -CH(CF )-CH-
)), 4.25 (m, 2H, CH -CH -O), 4.35 (t, 1H, CH -CH
CH(CF )-CH(NH )).
3
-CH
2
-CH-
(CF
3
)-CH(NH
2
)), 1.23 (t, CH
-CH(CF )-CH(NH
3
2
3
-
CH
2
3
2
3
2
3
(NH
2
3
2
3
2
-
3
2
D,L-2-Amino-3-trifluoromethylpentanoic Acid (2). Ethyl 2-amino-
3
-trifluoromethylpentanoate (200 mg, 0.94 mmol) was dissolved in 6
N HCl and the solution was refluxed for 1 day. The mixture was
evaporated to dryness in vacuo, and ethanol (5 mL) was added and
again evaporated to dryness in vacuo. The resulting solid was dissolved
in ethanol, neutralized with pyridine, and refrigerated over 2 days. The
precipitate was filtered and dried in vacuo to afford practically pure
1
amino acid (130 mg, 0.7 mmol, 74%). H NMR (D
2
O) δ 0.85 (t, 3H,
-CH -CH-
)-CH(NH )),
)), 3.81/3.92 (d, 1H, CH
O, TFA as reference) δ 7.1
CH
3
-CH
)-CH(NH
.81 (m, 1H, CH
-CH(CF )-CH(NH
2
-CH(CF
)), 1.5-1.8 (m, 1H, CH
-CH -CH(CF )-CH(NH
)). F NMR (D
3
)-CH(NH
2
)), 1.3-1.5 (m, 1H, CH
3
2
Protein Purification. The mouse dihydrofolate reductase (mDHFR)
encoded in pQE15 carries a 6× His tag sequence at the N-terminus.
mDHFR was purified by nickel affinity chromatography with stepwise
pH gradient elution under denaturing conditions according to the
(CF
3
2
3
-CH -CH(CF
2
3
2
2
3
2
3
2
3
-
1
9
CH
2
3
2
2
6
902 J. AM. CHEM. SOC. VOL. 125, NO. 23, 2003
9

Trifluoroisoleucine into Proteins in Vivo
A R T I C L E S
recommendations of the supplier (Qiagen). The eluted protein was
dialyzed (Spectra/Por membrane 1, MWCO ) 6-8 kDa) against
distilled water for 2 days and lyophilized. The purified protein was
subjected to matrix-assisted laser desorption ionization mass spectrom-
etry (MALDI-MS) and amino acid analysis.
concentration was measured using a Bio-Rad protein binding assay kit
with bovine serum albumin (BSA) as a standard.
Biological activity of mIL-2 was determined by a proliferation assay
using the IL-2 dependent H2-T cell line (ATCC).40 The serial 10-fold
dilutions of the renatured protein sample were incubated with H2-T
3
Tryptic Peptide Analysis. Purified protein (10 µL) in elution buffer
cells in 96-well plates (1 × 10 cells per well) overnight. Then cells
3
(
8 M urea, 100 mM NaH
with 90 µL 75 mM NH OAc and 2 µL of modified trypsin (Promega,
.2 µg/µL). The solution was incubated overnight at room temperature,
2
PO
4
, 10 mM Tris, pH ) 4.5) was mixed
were incubated with H-thymidine for 5 h and harvested onto filter
paper and analyzed on a Wallac (Gaithersburg, MD) 96-well counter.
4
0
Results
and quenched by addition of trifluoroacetic acid (pH < 4.0). The tryptic
digest was then subjected to sample cleanup by ZipTipC18 to provide 2
µL of purified sample solution. The MALDI matrix R-cyano-â-
Synthesis of Isoleucine Analogues. 3TFI was prepared from
1,1,1-trifluoro-2-butanone (4) as shown in Scheme 1. Treatment
3
hydroxycinnamic acid, (10 µL, 10 mg/mL in 50% CH CN) was added
of the fluoroketone with carboethoxymethylenetriphenylphos-
phorane gave a mixture of E and Z isomer of 3-trifluoromethyl-
to this solution, and 0.5 µL of the resulting solution was spotted directly
on the sample plate. Samples were analyzed on an Applied Biosystems
Voyager mass spectrometer operating in linear mode.
4
1,42
2
-pentenoic acid ethyl ester (94:6),
which was converted to
the saturated ester by low-pressure hydrogenation. Hydrolysis
to the acid (5), followed treatment with thionyl chloride and
bromine, and finally quenching with ethanol afforded ethyl
IleRS Cloning, Expression and Purification. The isoleucyl-tRNA
synthetase (IleRS) gene was cloned directly from E. coli genomic DNA
with two flanking primers encoding the desired restriction sites BamHI
and SacI (primer 1: 5′-CAA ACC GAA ATA CGG ATC CGA GAA
TCT GAT G-3′; primer 2: 5′-CTG TTG AAC AGA TCG ATT GAG
CTC ATC AGG CAA ACT TAC-3′). The resulting 3000 base-pair
DNA fragment was gel-purified, digested with BamHI and SacI and
ligated into the expression plasmid pQE30 (Qiagen) to yield pQE-ileS.
The integrity of the cloned gene was confirmed by DNA sequencing.
The cloned synthetase carried the N-terminal leader sequence MRG-
SHHHHHHGSENL. pQE-ileS was transformed into XL-1 (Stratagene)
to yield the expression strain. Protein expression was induced at
OD600 ) 0.6 with 1 mM IPTG. After 3 h, the cells were harvested.
The enzyme was purified using Ni-NTA agarose resin under native
conditions according to the manufacturer’s instructions (Qiagen). Protein
was stored in Buffer A (50 mM Tris-HCl, 1 mM DTT)/50% glycerol.
Aliquots were flash frozen and stored at -80 °C. The concentration of
the purified enzyme was determined by absorbance at 280 nm under
denaturing conditions.
2
-bromo-3-trifluoromethylpentanoate. Walborsky et al. found
that ammonolysis of ethyl 2-bromo-4,4,4-trifluorobutyrate gave
3-amino-4,4,4-trifluorobutyramide instead of the desired 2-ami-
4
3
no isomer. By using azide ion to decrease the tendency for
elimination and to favor direct displacement of bromide,
44
Loncirni et al. obtained ethyl 2-azido-3,3,3-trifluorobutyrate.
We used a similar procedure to prepare ethyl 2-azido-3-tri-
fluoromethylpentanoate (6), which after catalytic hydrogenation
and hydrolysis provided 3TFI in an overall yield of 5% based
on 4. The resulting mixture of diastereomers of 3TFI was used
in studies of incorporation into proteins in vivo.
45
5TFI was prepared according to Muller as shown in Scheme
2. One pair of enantiomers, corresponding to 5,5,5-trifluoro-
DL-isoleucine, was separated by fractional crystallization. We
use the designation 5TFI to refer to the enantiomeric pair.
Incorporation of 5TFI into Protein in Vivo. We used E.
coli strain AI-IQ[pQE15] to evaluate production of the test
protein mDHFR in cultures supplemented with 3TFI and with
ATP-PP
of isoleucine analogues by IleRS was carried out via an ATP-PP
exchange assay as described by Schimmel and co-workers.38 The assay
buffer contained 30 mM HEPES, pH 7.4, 10 MgCl , 1 mM DTT, 2
(0.5 TBq/mol). The enzyme concentra-
i
Exchange Assay. Measurement of the rates of activation
i
2
5
TFI, respectively. The parent isoleucine auxotrophic strain AI
32
mM ATP and 2 mM [ P]-PP
i
was generated from E. coli strain BL21(DE3) by P1-mediated
tion was 100 nM. The amino acid concentration varied depending on
the activity of enzyme toward the substrate (1: 10-500 µM; 2, 3: 30-
transduction using a donor strain (BW 6159), which had been
3
7
transposed by Tn10 at the ileD gene region; ileD is essential
for the endogenous synthesis of isoleucine and valine.46 The
AI-IQ[pQE15] culture was grown to OD600 ) 1.0 in supple-
mented M9 minimal medium, then shifted to a medium
containing either: (a) the analogue of interest, (b) Ile (positive
control), or (c) no Ile or analogue (negative control). Protein
synthesis was induced by addition of IPTG. The results of
expression of mDHFR are shown by SDS-PAGE analysis of
whole cell lysates (Figure 1). No expression of mDHFR was
detected in negative control cultures or in cultures containing
1
0,000 µM). Aliquots (20 µL) of reaction mixture were quenched into
5
00 µL quench buffer (200 mM PP , 7% w/v HClO and 3% activated
i
4
charcoal). The charcoal was washed twice with wash buffer (10 mM
PP , 0.5% HClO ) and counted by liquid scintillation methods. Kinetic
i
4
parameters were obtained by nonlinear regression analysis.
Incorporation of 5TFI into Murine Interleukin-2. Murine inter-
leukin-2 (mIL-2) was cloned into pQE31 (Qiagen) by PCR cloning
through restriction sites BamHI and SalI. The cloned gene was
confirmed by DNA sequencing. The resulting plasmid pQE-mIL2 was
transformed into AI-IQ to yield expression strain AI-IQ[pQE-mIL2].
Expression was performed essentially as described above. Tryptic
peptide digestion was employed to confirm the incorporation of 3 into
mIL-2. Protein was first purified using Ni-NTA agarose resin under
denaturing conditions according to the manufacturer’s instructions
3
TFI (500 mg/L). In contrast, synthesis of mDHFR is evident
in cultures supplemented with Ile (20 mg/L) or with 5TFI (500
mg/L). We also explored a modified expression host (AI-IQ-
[pQE15-IRS]), in which the isoleucyl-tRNA synthetase was
overexpressed. Amino acid activation assays on whole cell
lysates indicated that the IleRS activity of the modified host
was elevated approximately 6-fold as compared to the wild-
type host. Nevertheless, qualitatively similar results were
(Qiagen). His-tag purified fractions were pooled and diluted in 8 M
urea to a final protein concentration of less than 25 mg/L. The protein
3
9
was refolded according to a literature procedure. The diluted protein
solution was stepwise dialyzed against the following solutions: (1) 6
M urea, 2.5 mM DTT, 2.5 mM â-mercaptoethanol (â-ME), 10 mM
Tris‚HCl, pH ) 7.4; (2) 4 M urea, 5 mM â-ME, 10 mM Tris‚HCl,
pH ) 7.4; (3) 2 M urea, 5 mM â-ME, 10 mM Tris‚HCl, pH ) 7.4; (4)
(
43) Walborsky, H. M. B., M.; Loncrini, D. F. J. Am. Chem. Soc. 1955, 77,
H
2
O, 5 mM â-ME; (5) H
2
O, 1 mM DTT; (6) HEPES-buffered saline,
3637.
(
44) Loncrini, D. F.; Walborsky, H. M. J. Med. Chem. 1964, 7, 9.
45) Muller, N. J. Fluorine Chem. 1987, 36, 163-170.
1
mM DTT; (7) 50 mM HEPES, pH ) 7.4, 1 mM DTT. The renatured
(
protein was concentrated via ultrafiltration and stored at -80 °C. Protein
(46) Im, S. W. K. P., J. J. Bacteriol. 1971, 106, 784-790.
J. AM. CHEM. SOC.
9
VOL. 125, NO. 23, 2003 6903

A R T I C L E S
Wang et al.
Figure 1. SDS-PAGE analysis of mDHFR synthesized by AF-IQ[pQE15]
or AF-IQ[pQE-IleRS]. Expression conditions are noted above each lane.
The lane in the middle is the molecular weight marker. pQE-IleRS is the
plasmid derived from pQE15 (Qiagen), with insertion of the IleRS
overexpression cassette at the NheI site.
Table 1. Protein Yield and Extent of Isoleucine Replacement in
MDHFR
Figure 3. Growth rate of E. coli strain AF-IQ[pQE15] as a function of
the concentration of 5TFI in the culture medium.
5TFI replacement (%)
Table 2. Kinetic Parameters for Activation of 1 and 3 by
E. coli IleRS
protein
yield (mg/L)
mass (Da)
AAA
MALDI
mDHFR-WT
mDHFR-5TFI
30
20
24063
24818
amino acid
K
m
(µM)
6.82
30.52
k
cat (1/s)
k
cat/K
m
(rel)
93.2
99.8
1
3
3.0
0.1
1
1/134
Effect of 5TFI on Cell Growth. We also determined the
effect of 5TFI on the growth of the E. coli host strain prior to
induction. Isoleucine is required for growth of E. coli strain
AI. In cultures supplemented with 0.2 mM isoleucine, the rate
of cell growth is dependent on the concentration of 5TFI. At
an analogue concentration of 4 mM, cell growth is completely
inhibited (Figure 3). The dose dependent toxicity of 5TFI
presumably arises from incorporation of the fluorinated amino
acid into cellular proteins.
Amino Acid Activation in Vitro. The rates of ATP-
dependent activation of isoleucine and isoleucine analogues 2
and 3 by E. coli isoleucyl-tRNA synthetase (IleRS) were
examined by the ATP-PPi exchange assay. The kinetic
parameters are reported in Table 2. The values of Km and kcat
for Ile are in the range of values reported in the literature.4
Because 3TFI causes no exchange of PPi above background
level, no kinetic parameters were obtained. Comparison of the
specificity constants (kcat/Km) obtained for isoleucine (0.44
Figure 2. MALDI analysis of mDHFR following trypsin digestion. A
fragment of sequence VDMVWIVGGSSVYQEAMNQPGHLRLFVTR,
yields the spectra shown. (a) fragment digested from mDHFR-WT; (b)
fragment digested from mDHFR-5TFI.
7-50
obtained from incorporation experiments on modified and wild-
type strains (Figure 1).
-1
-1
-3 -1
-1
s µM ) and for 5TFI (3.28 × 10 s µM ) show that the
fluorinated analogue is activated roughly 100-fold less efficiently
than the natural substrate.
After purification by nickel-affinity chromatography, the yield
of protein (mDHFR-5TFI) expressed in 5TFI-supplemented
medium was 20 mg/L, approximately 60% of that isolated from
the positive control medium (Table 1). Amino acid analysis
showed the extent of substitution of 5TFI for Ile to be 93%.
Purified mDHFR was also subjected to MALDI-TOF mass
spectrometry. The difference in molar mass between mDHFR-
5
TFI in Murine Interleukin-2. 5TFI was efficiently incor-
porated into mIL-2, as shown by SDS-PAGE and MALDI
analysis of tryptic peptide fragments (Figure 4). Amino acid
analysis indicated that 85% of isoleucine was replaced by 3.
Refolded mIL-2 samples were analyzed for activity by measur-
3
ing the rate of [ H]-thymidine incorporation in the mIL-2
5TFI and mDHFR-wt is 755 Da (Table 1), consistent with the
dependent cell line H2-T stimulated by varying concentrations
of the wild-type and fluorinated proteins. Dose-response curves
were similar for both forms of the protein, as shown in Figure
5. As shown in Table 3, the concentration of fluorinated protein
that elicits 50% of the maximal proliferative response is 3.87
ng/mL, about 30% higher than that of wild-type mIL-2
(EC50 ) 2.70 ng/mL). The maximal responses are equivalent
for the fluorinated and wild-type cytokines.
fact that there are 14 isoleucine residues in mDHFR and with
the mass difference of 54 Da between 5TFI and Ile.
Incorporation of 5TFI was also confirmed by tryptic peptide
analysis of purified mDHFR. Figure 2a shows a representative
tryptic fragment of mDHFR-wt with a mass of 2674.15 Da.
This fragment is assigned to residues 124-147 (Figure 2a) and
contains 1 of the 14 isoleucine residues of mDHFR. For
mDHFR-5TFI, the mass of this fragment is shifted to 2728.04
Da (Figure 2b). The additional mass (53.88 Dalton) corresponds
to the mass difference between 5TFI and Ile. The original signal
has disappeared, consistent with near quantitative substitution
of Ile by 5TFI.
(
47) Pope, A. J.; Lapointe, J.; Mensah, L.; Benson, N.; Brown, M. J. B.; Moore,
K. J. J. Biol. Chem. 1998, 273, 31 680-31 690.
(48) Jakubowski, H.; Fersht, A. R. Nucl. Acids Res. 1981, 9, 3105-3117.
(
49) Moe, J. G.; Piszkiewicz, D. Biochemistry 1979, 18, 2804-2810.
50) Busiello, V.; Digirolamo, M.; Demarco, C. Biochimica Et Biophysica Acta
1979, 561, 206-214.
(
6904 J. AM. CHEM. SOC.
9
VOL. 125, NO. 23, 2003

Trifluoroisoleucine into Proteins in Vivo
A R T I C L E S
Figure 6. Crystal structure of hIL-2 (PDB: 1M47). The identified sites of
interaction of hIL-2 with its receptor are shown in blue. The five conserved
isoleucine residues are represented by ball-and-stick models.
Figure 4. MALDI analysis of mIL-2 following trypsin digestion. A
fragment of sequence SFQLEDAENFISNIR yields the spectra shown. (a)
fragment digested from wild-type mIL-2; (b) fragment digested from mIL-2
containing 5TFI.
us initially to engineer an E. coli host strain with enhanced IleRS
activity, but we found that 3TFI does not support measurable
protein synthesis even with overexpression of IleRS.
The results of in vitro activation assays are consistent with
the in vivo results. 5TFI shows modest activity (kcat/Km 134-
fold less than that of isoleucine), whereas only background levels
of pyrophosphate exchange were observed in experiments with
3
TFI. The origins of the difference in activity are under
investigation.
The high fidelity of IleRS is due in part to the editing activity
of the synthetase. A double-sieve mechanism has been proposed
5
4
55
by Fersht and co-workers and verified in structural studies.
According to this model, IleRS has two sites for amino acid
discrimination: the synthetic site serves as a coarse sieve and
precludes amino acids larger than Ile from binding; the editing
site serves as a fine sieve that removes misactivated amino acids
smaller than Ile. Our in vitro results suggest that the synthetic
site binds 5TFI, but not 3TFI. Results from in vivo incorporation
studies confirm that 5TFI is not edited, consistent with the
enhanced steric bulk of the fluorinated side chain.
Figure 5. Proliferative response of IL-2 dependent H2-T cells to
recombinant mIL-2 proteins: wild-type mIL-2 (9); mIL-2 containing 5TFI
(O). The results are averages from triplicate experiments.
Table 3. Proliferative Response of H2-T Cells
protein
EC50a (ng/mL)
The extent to which globular proteins can tolerate incorpora-
tion of noncanonical amino acids without loss of function
remains largely unexplored. Interleukin-2 (IL-2) is one of the
major cytokines mediating the immune system, with the
responsibility of T cell clonal expansion after antigen recogni-
mIL2-wt
mIL2-5TFI
2.70 ( 0.31
3.87 ( 0.22
a
Concentration of protein eliciting 50% of maximal response as measured
by proliferation assay. Results from triplicate determinations.
5
6
tion. The crystal structure of human IL-2, a four-helix bundle
Discussion
57
protein, is shown in Figure 6. The sites of interaction with
5
8
the IL-2 receptor are shown in blue. All 5 isoleucine residues
in mIL-2 are conserved in hIL-2 and are located in the helical
core structure (Figure 6). Amino acid analysis indicated 85%
substitution of Ile by 5TFI in the sample of mIL-2 prepared
here; thus the fraction of wild-type cytokine in the fluorinated
sample is negligible. The proliferative response elicited by
fluorinated mIL-2 indicates that the fluorinated protein must
fold into an authentic, native structure. The generality of this
Prior to the start of this work, it was known that isoleucine
analogues 2-amino-3-methylthiobutyric acid, 2-amino-3-meth-
5
1
52
ylaminobutyric acid, O-methylthreonine, and â-methylnor-
leucine53 inhibit the growth of E. coli at millimolar concentra-
tions, suggesting that these analogues might serve as substrates
for the isoleucyl-tRNA synthetase and find their way into
cellular proteins. Our results show clearly that 5TFI serves as
an effective isoleucine surrogate in the translation step of protein
biosynthesis in E. coli. 3TFI behaves in a strikingly different
manner: no evidence for incorporation has been found. The
fact that we did not observe incorporation of 3TFI prompted
(
54) Fersht, A. R.; Dingwall, C. Biochemistry 1979, 18, 1238-1245.
(55) Nureki, O.; Vassylyev, D. G.; Tateno, M.; Shimada, A.; Nakama, T.; Fukai,
S.; Konno, M.; Hendrickson, T. L.; Schimmel, P.; Yokoyama, S. Science
1998, 280, 578-582.
(
56) Abbas, A. K.; Lichtman, A. H.; Pober, J. S. Cellular and Molecular
Immunology; W. B. Saunders Company: Philadelphia, 2000.
(
51) McCord, T. J.; Foyt, D. C.; Kirkpatrick, J. L.; Davis, A. L. J. Med. Chem.
1
967, 10, 353-355.
(57) Arkin, M. A.; Randal, M.; Delano, W. L.; Hyde, J.; Luong, T. N.; Oslob,
J. D.; Raphael, D. R.; Taylor, L.; Wang, J.; Wells, J. A.; Mcdowell, R. S.;
Braisted, A. C. Proc. Natl. Acad. Sci. U.S.A. 2003, 100, 1603-1606.
(58) Mott, H. R.; Campbell, I. D. Curr. Opin. Struct. Biol. 1995, 5, 114-121.
(
52) Mourad, G.; King, J. Plant Physiol. 1995, 107, 43-52.
(
53) Masaki, S.; Masahiko, K.; Chibata, I. Agric. Biol. Chem. 1985, 49, 1889-
1
990.
J. AM. CHEM. SOC.
9
VOL. 125, NO. 23, 2003 6905

A R T I C L E S
Wang et al.
result remains to be determined; however, these observations
coupled with our previous reports25,30,32 suggest that at least
some protein domains will tolerate side-chain fluorination
without loss of function.
from David Baltimore’s laboratory at Caltech and Dr. Susan
Kovats at City of Hope for assistance on the cell-proliferation
assay. We thank Dr. Osamu Nureki at the University of Tokyo
for providing coordinates of the crystal structure of IleRS
complexed with isoleucine, and Kent Kirshenbaum for helpful
discussions.
Acknowledgment. This work was supported by NIH grant
R01-GM62523, and by the NSF Center for the Science and
Engineering of Materials at Caltech. We are grateful to Lili Yang
JA0298287
6906 J. AM. CHEM. SOC.
9
VOL. 125, NO. 23, 2003