Genetic introduction of a diketone-containing amino acid into proteins

Article abstract of DOI:10.1016/j.bmcl.2006.07.094

An orthogonal tRNA/aminoacyl-tRNA synthetase pair was evolved that makes possible the site-specific incorporation of an unnatural amino acid bearing a β-diketone side chain into proteins in Escherichia coli with high translational efficiency and fidelity.

Full text of DOI:10.1016/j.bmcl.2006.07.094

Bioorganic & Medicinal Chemistry Letters 16 (2006) 5356–5359
Genetic introduction of a diketone-containing
amino acid into proteins
Huaqiang Zeng, Jianming Xie and Peter G. Schultz*
Department of Chemistry and The Skaggs Institute for Chemical Biology, The Scripps Research Institute,
10550 North Torrey Pines Road, La Jolla, CA 92037, USA
Received 22 June 2006; revised 21 July 2006; accepted 24 July 2006
Available online 24 August 2006
Abstract—An orthogonal tRNA/aminoacyl-tRNA synthetase pair was evolved that makes possible the site-specific incorporation of
an unnatural amino acid bearing a b-diketone side chain into proteins in Escherichia coli with high translational efficiency and fidel-
ity. Proteins containing this unnatural amino acid can be efficiently and selectively modified with hydroxylamine derivatives of fluo-
rophores and other biophysical probes.
Ó 2006 Elsevier Ltd. All rights reserved.
Recently, we showed that by adding a unique amber sup-
pressor tRNA (mutRNA
methyl ester-protecting groups (Scheme 1). The second
carbonyl group was added using potassium tert-butox-
ide in the mixed solvent, 2:3 [v/v] methyl acetate/
THF.14 Removal of t-Boc group with TFA, followed
by alkaline hydrolysis, afforded the desired diketone-
containing unnatural amino acid 1 in an overall yield
of 40%.
Tyr
CUA
)/aminoacyl-tRNA synthe-
tase (MjTyrRS) to the translational machinery of
Escherichia coli, unnatural amino acids can be incorpo-
rated into proteins in response to nonsense and frameshift
codons with high translational fidelity and efficiency.1–5
This approach has been used to genetically encode over
30 unnatural amino acids including glycosylated, photo-
reactive, metal binding, and heavy atom-containing ami-
no acids.6–8 In addition, a series of chemically reactive
amino acids with ketone-, azide-, and acetylene-contain-
ing side chains have been selectively incorporated into
proteins. These mutant proteins can be subsequently
modified by nonpeptidic molecules with useful biological
and/or physical properties (e.g., polyethylene glycol,
biotin, glycomimetics, fluorophores, cross-linking agents,
and cytotoxic molecules).9–13 Because of the unique reac-
tivity of these amino acid side chains, such bioconjuga-
tion reactions are highly selective. Herein, we extend
this methodology to the genetic incorporation of dike-
tone-containing amino acid 1 in E. coli. The b-diketone
group can react to form stable complexes with a number
of functional groups including alkoxyamines, hydrazides,
and primary amines.
We previously generated an amber suppressor tRNA/
Tyr
CUA
aminoacyl-tRNA synthetase pair (mutRNA
-
MjTyrRS) derived from the tRNATyr/TyrRS pair of
Methanococcus jannaschii, which has been used to selec-
tively and efficiently incorporate a large number of
unnatural amino acids in E. coli in response to the
TAG codon.9,10,12,15 On the basis of a crystal structure
of the M. jannaschii TyrRS-tRNA(Tyr)-L-tyrosine com-
plex,16 six residues (Tyr32, Leu65, Phe108, Gln109, Asp158
and Leu162) in the tyrosine-binding site of M. jannaschii
,
O
O
O
O
O
O
iv, v
H2N COOH
iii
i, ii
t-Boc
t-Boc
N
COOCH3
1c
H2N COOH
N
COOCH3
1b
H
H
Diketone 1 was synthesized from p-acetyl-L-phenylala-
nine containing tert-butyloxycarbonyl (t-Boc) and
1a
1
Scheme 1. Reagents and conditions: (i) SOCl2, anhydrous MeOH, 1 h
at rt, 90%; (ii) t-(Boc)2O, Et3N, CH2Cl2, 2 h at rt, 85%; (iii) potassium
tert-butoxide, 40% methyl acetate in THF, 1 h at rt, 60%; (iv) TFA,
CH2Cl2, 5 h at rt, 92%; (v) NaOH, 30% H2O in MeOH, 6 h at rt,
followed by addition of aqueous concentrated HCl, >90%.
Keywords: Bioorganic chemistry; Protein modification; In vivo selec-
tion; Aminoacyl-tRNA synthetase; Diketone-containing amino acid.
*
Corresponding author. E-mail: schultz@scripps.edu
0960-894X/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2006.07.094

H. Zeng et al. / Bioorg. Med. Chem. Lett. 16 (2006) 5356–5359
5357
TyrRS were randomly mutated. The mutant MjTyrRS
a
MTSVDNK7INKEQQNAFYEILHLPNLNEEQ
RDAFIQSLKDDPSQSANLLAEAKKLNDAQ
APKGSHHHHHH
(mutMjTyrRS) library (>109 clones) was then passed
through three rounds of positive selection in the pres-
ence of 1 mM unnatural amino acid 1 (based on the sup-
pression of an amber stop codon in the chloramphenicol
acetyl transferase (CAT) gene), alternated with two
rounds of negative selection in the absence of 1 (based
on the suppression of three amber stop codons intro-
duced into the toxic barnase gene).9 One clone emerged
whose survival in chloramphenicol was dependent on
the presence of 1; this mutant MjTyrRS supported cell
growth in 120 lg mLꢀ1 chloramphenicol in the presence
of 1, and up to 10 lg mLꢀ1 chloramphenicol in the ab-
sence of 1. This result suggests that the mutant synthe-
tase has higher specificity for 1 than for endogenous
amino acids. Sequencing revealed the following muta-
tions: Tyr32 ! Gly32, Asp158 ! Gly158, Phe65 ! Val65,
Phe108 ! Thr108, and Leu162 ! Ser162. The Tyr32 and
Asp158 mutations remove the two hydrogen bonds to
tyrosine in the wildtype (wt) synthetase, consistent with
the decreased activity of the enzyme toward its natural
substrate.
1
2
3
4
b
Figure 1. (a) The sequence of the Z-domain protein. The codon for
Lys7 was mutated to TAG in order to introduce amino acied 1. (b)
Gelcode Blue-stained SDS–PAGE analysis of Lys7 ! TAG7 Z-domain
protein expressed under different conditions. Lane 1: molecular mass
marker; lane 2: expression with wt MjTyrRS; lane 3: expression with
mutMjTyrRS in the presence of 1; lane 4: expression with mutMjTyr-
RS in the absence of 1. wt = wildtype.
To confirm that the observed phenotype is caused
by the site-specific incorporation of
1 by the
mutRNACTUyAr –mutMjTyrRS pair, an amber stop codon
was substituted at a permissive site (Lys7) in the gene
encoding the Z-domain protein17 fused to a C-termi-
7998
7909
O
7867
O
5000
nal hexameric His tag. Cells transformed with
, mutMjTyrRS, and the mutant Z-do-
Tyr
CUA
mutRNA
4000
3000
2000
main gene were grown in the presence or absence of
1 mM 1 in minimal medium containing 1% glycerol
and 0.3 mM leucine (GMML medium). The expres-
sion of the Lys7 ! TAG7 Z-domain protein by
mutMjTyrRS was induced by the addition of 1 mM
isopropyl-b-D-thiogalactopyranoside (IPTG) and pro-
tein was purified by Ni2+ affinity chromatography.
Analysis by SDS–PAGE revealed that the expression
of mutant protein was strongly dependent on 1—the
yield of purified mutant protein was 1.6 mg/liter of
culture in the presence of 1, and insignificant in the
absence of 1 (Fig. 1), consistent with a high fidelity
for the incorporation of 1 into proteins by mutMjTyr-
RS. For comparison, the yield of the wt Z-domain
protein using wt MjTyrRS that charges only endoge-
nous tyrosine in response to the amber stop codon
was 3.0 mg/liter of culture under the identical condi-
tions. Additional evidence for the site-specific incorpora-
tion of 1 was obtained by matrix-assisted laser desorption/
ionization time-of-flight mass spectrometry (MALDI-
TOF MS).18 In addition to the observation of an experi-
mental average mass of 7997 Da (MTheoretical = 7997 Da)
for the intact mutant protein, a major peak corresponding
to the protein without the first methionine moiety
(MExperimental = 7866 Da, MTheoretical = 7866 Da) was also
8040
1000
0
7000
7500
8000
Mass (m/z)
8500
9000
Figure 2. MALDI-TOF analysis of the mutant Z-domain protein
containing 1.18 The observed masses of intact mutant protein in its
protonated and acetylated forms are 7998 and 8040 Da, respectively.
The observed masses of mutant protein without the first methionine in
its protonated and acetylated forms are 7867 and 7909 Da, respective-
ly. All of these experimental data are in excellent agreement with
theoretically calculated values (see text for more detail).
The possibility of using the diketone moiety as a chem-
ical handle for the selective modification of protein with
exogenous agents was then tested by labeling diketone-
containing protein with a biotin hydroxylamine deriva-
tive (Fig. 3a, MW = 331.39, purchased from Molecular
Probes).9 The purified mutant and wt Z-domain proteins
(ꢁ0.1 mM) were treated with 1.0 mM biotin hydroxyl-
amine in PBS buffer (20 mM sodium phosphate,
150 mM NaCl) at pH 4.0 at 25 °C for 12 h. After dialysis
against distilled water to remove excess biotin hydroxyl-
amine, the proteins were analyzed by MALDI-TOF MS.
detected (Fig. 2). The other two peaks (MExperimental
=
8039 Da and MExperimental = 7908 Da) correspond to the
mutant proteins with or without the first methionine in
their acetylated forms (MTheoretical = 8039 Da and
MTheoretical = 7908 Da), respectively.19 These results con-
Experimental average masses of 8315 Da (MTheoretical
8310 Da for biotin-labeled intact mutant protein),
8182 Da (MTheoretical = 8179 Da for biotin-labeled
=
firm a high fidelity for the incorporation of 1 by
Tyr
mutRNA
/mutMjTyrRS pair.
CUA

5358
H. Zeng et al. / Bioorg. Med. Chem. Lett. 16 (2006) 5356–5359
To demonstrate the generality of this approach, the
a
b
-
-
SO3
SO3
H
H
labeling of Z-domain protein with a fluorophore
hydroxylamine derivative was also carried out. The
purified mutant and wt Z-domain proteins
(ꢁ0.05 mM) were treated with 0.5 mM Alexa Fluor
488 hydroxylamine derivative (Fig. 3b, Molecular
Probes) in PBS buffer at pH 4.0 at 25 °C for 12 h
(Fig. 3b). After conjugation, proteins were dialyzed
against distilled water to remove excess fluorophore
and then analyzed by SDS–PAGE. The gel was first im-
aged with a Phosphoimager21 (Fig. 3d) and then stained
with Gelcode Blue (Fig. 3c). The band for mutant Z-do-
main shows a fluorescent signal, whereas no fluorescence
at the same position can be detected for either the wt Z-
domain or dye molecules alone, indicating that Alexa
Fluor 488 hydroxylamine derivative was selectively
coupled to the diketone group in the presence of the
other amino acid side chains in the protein. The labeled
protein was confirmed by MALDI-TOF MS
(MExperimental = 8678 Da and MTheoretical = 8668 Da
for the dye-labeled intact protein) to be the product
formed between one molecule of Alexa Fluor 488
hydroxylamine and one molecule of mutant Z-domain.
The labeling efficiency was >85% as estimated by
MALDI-TOF analysis (Fig. 4d).20
+H2N
O
N
O
NH2
S
H
NH
-OOC
H
(CH2)4
C
O
6
O
NH(CH2)5NH
C
O
CH2ONH2
5
NHNH
C CH2ONH2
O
c
Marker
wt
Mutant
Dye only
d
Figure 3. In vitro labeling of the mutant Z-domain protein containing
1 with hydroxylamine derivatives of (a) biotin and (b) Alexa Fluor 488.
(c) Gelcode Blue-stained SDS/PAGE analysis of wt and mutant
Z-domain proteins treated with Alexa Fluor 488 hydroxylamine. (d)
Fluorescence imaging of wt and mutant Z-domain proteins treated
with Alexa Fluor 488 hydroxylamine. wt = wildtype.
The present study demonstrates the site-specific incor-
poration of the b-diketone moiety into proteins in vivo
with high fidelity and good efficiency by means of an
amber suppressor tRNA/aminoacyl-tRNA synthetase
pair. The b-diketone can serve as a unique chemical
handle for the efficient and selective modification of
target proteins with a host of exogenous alkoxyamine
derivatives in vitro. Additionally, the b-diketone group
may be able to participate in carbon–carbon bond
forming Michael reactions.22,23 Moreover, an enamine
adduct stabilized by a six-membered intramolecular
hydrogen bond can be formed between the b-diketone
and a primary amine.24 Consistent with this observa-
tion, we found that, while an imine adduct formed be-
tween an aryl monoketone and butyl amine readily
undergoes hydrolysis at pH 4.0–10.4, the enamine de-
rived from the b-diketone is highly resilient toward
hydrolysis in the same pH range.25 Consequently, by
placing a diketone-containing unnatural amino acid
immediately adjacent to a lysine residue situated in a
favorable hydrophobic environment, one may be able
to form intermolecular or intramolecular protein
crosslinks.
mutant protein without the first methionine residue),
and 8225 Da (MTheoretical = 8221 Da for biotin-labeled
mutant protein without the first methionine residue in
its acetylated form) were found, confirming that biotin
hydroxylamine reacted with the mutant Z-domain pro-
teins in a molar ratio of 1:1. As expected, no labeling
products were detected for wt Z-domain protein, indi-
cating that the labeling reaction occurred only between
the hydroxylamine and the diketone group, but not
any existing functional groups in the wt protein. Based
on the integration ratio of the peaks in the MALDI-
TOF spectrum, the labeling efficiency was estimated to
be >85% (Fig. 4c).20
a
b
c
d
Labeled
Labeled with
Labeled
Labeled with
Alexa Fluor 488
Unlabeled
with biotin Alexa Fluor 488 with biotin
Unlabeled
Acknowledgment
This work was supported by a Department of Energy
Grant DE-FG03-00ER46051 to P.G.S.
8000
8000
8000
8000
10000
10000
10000
10000
References and notes
Figure 4. MALDI-TOF analysis18 of doping experiments with com-
parable amounts of unlabeled diketone-containing Z-domain protein
and its corresponding labeled versions with either (a) biotin or (b)
Alexa Fluor 488, and of the diketone-containing Z-domain protein
labeled with (c) biotin and (d) Alexa Fluor 488.20
1. Wang, L.; Brock, A.; Herberich, B.; Schultz, P. G. Science
2001, 292, 498.
2. Chin, J. W.; Cropp, T. A.; Anderson, J. C.; Mukherji, M.;
Zhang, Z. W.; Schultz, P. G. Science 2003, 301, 964.

H. Zeng et al. / Bioorg. Med. Chem. Lett. 16 (2006) 5356–5359
5359
3. Anderson, J. C.; Wu, N.; Santoro, S. W.; Lakshman, V.;
King, D. S.; Schultz, P. G. Proc. Natl. Acad. Sci. U.S.A.
2004, 101, 7566.
4. Zhang, Z. W.; Alfonta, L.; Tian, F.; Bursulaya, B.; Uryu,
S.; King, D. S.; Schultz, P. G. Proc. Natl. Acad. Sci.
U.S.A. 2004, 101, 8882.
5. Zhang, Z. W.; Gildersleeve, J.; Yang, Y. Y.; Xu, R.; Loo,
J. A.; Uryu, S.; Wong, C. H.; Schultz, P. G. Science 2004,
303, 371.
6. Wang, L.; Schultz, P. G. Chem. Commun. 2002, 1.
7. Wang, L.; Schultz, P. G. Angew. Chem. Int. Ed. 2005, 44,
34.
19. Although the enzyme that catalyzes acetylation of Z-
domain protein is not known, it is very likely that protein
acetylation of the Z-domain protein is mediated by known
or unknown bacterial N-acetyltransferases at either the N-
terminus or one of five internal lysine residues found in
Z-domain protein. For the recent description of bacterial
N-acetyltransferases, see: (a) Brooke, E. W.; Davies, S. G.;
Mulvaney, A. W.; Pompeo, F.; Sim, E.; Vickers, R. J.
Bioorg. Med. Chem. 2003, 11, 1227; (b) Vetting, M. W.; S
de Carvalho, L. P.; Yu, M.; Hegde, S. S.; Magnet, S.;
Roderick, S. L.; Blanchard, J. S. Arch. Biochem. Biophys.
2005, 433, 212.
8. Xie, J.; Schultz, P. G. Curr. Opin. Chem. Biol. 2005, 9, 548.
9. Wang, L.; Zhang, Z.; Brock, A.; Schultz, P. G. Proc. Natl.
Acad. Sci. U.S.A. 2003, 100, 56.
10. Zhang, Z.; Smith, B. A. C.; Wang, L.; Brock, A.; Cho, C.;
Schultz, P. G. Biochemistry 2003, 42, 6735.
11. Liu, H.; Wang, L.; Brock, A.; Wong, C.-H.; Schultz, P. G.
J. Am. Chem. Soc. 2003, 125, 1702.
12. Deiters, A.; Cropp, T. A.; Mukherji, M.; Chin, J. W.;
Anderson, J. C.; Schultz, P. G. J. Am. Chem. Soc. 2003,
125, 11782.
20. Doping experiments with comparable amounts of dike-
tone-containing Z-domain protein and its corresponding
labeled versions with either biotin or Alexa Fluor 488
demonstrated that these proteins in either unlabeled or
labeled forms have comparable ionization efficiencies
under mass detection conditions. In calculating the label-
ing efficiency, the small peak areas from 7867 to 8040 Da
were taken as the upper-limit value for unlabeled dike-
tone-containing proteins. The integration of the corre-
sponding peak areas from 8182 to 8357 Da for biotin-
labeled proteins (Fig. 4c) and from 8540 to 8715 Da
(Fig. 4d) for dye-labeled proteins in their respective
MALDI-TOF spectra suggests labeling efficiencies of
greater than 85% for conjugation reactions involving
biotin and Alexa Fluor 488 molecules. A correction factor
of 0.8 obtained by integration of their corresponding peak
areas in Figure 4b has been taken into account when
calculating the labeling efficiency involving Alexa Fluor
488 molecule.
21. The fluorescence image was recorded on a Molecular
Dynamics Storm 860 Phosphoimager (Molecular Dynam-
ics, Sunnyvale, CA) by chemi-fluorescence/blue-excited
scans at 450 nm, both excitation and emission bandpass of
4 nm, a photomultiplier tube voltage of 750 V, and a scan
rate of 1 nm/s. 0.1–0.3 nanomoles of each protein were used.
22. Carlqvist, P.; Svedendahl, M.; Branneby, C.; Hult, K.;
Brinck, T.; Berglund, P. ChemBioChem 2005, 6, 331.
23. Svedendahl, M.; Hult, K.; Berglund, P. J. Am. Chem. Soc.
2005, 127, 17988.
13. Tsao, M. L.; Tian, F.; Schultz, P. G. ChemBioChem 2005,
6, 2147.
14. To a solution of 1b (3.85 g, 12.0 mmol) in methyl acetate
(24 mL) was added dropwise 36 mL of 1 M potassium
tert-butoxide in THF at room temperature, and the
solution was stirred for 45 min. After the removal of
THF, CH2Cl2 (300 mL) was added, which was washed
twice with distilled water, once with brine and dried over
anhydrous Na2SO4. The solvent was removed and the
resulting solid was subjected to the flash chromatography
on a silical gel (30:70 [v/v] ethyl acetate/hexane) to afford
1
pure product 1c (2.60 g, 60%) as a white solid. H NMR
(399 MHz, CDCl3, 298 K): d 7.80 (d, J = 8.4 Hz, 2H), 7.21
(d, J = 8.0 Hz, 2H), 6.26 (s, 2H), 4.99 (m, 1H), 4.61 (m,
1H), 3.72 (s, 3H), 3.38–2.90 (m, 2H), 2.20 (s, 3H), 1.42 (s,
9H). Exact mass m/z without Boc group calcd for
C14H17NO4 263.1, found (LC/MS) 263.1.
15. Alfonta, L.; Zhang, Z.; Uryu, S.; Loo, J. A.; Schultz, P. G.
J. Am. Chem. Soc. 2003, 125, 14662.
16. Kobayashi, T.; Nureki, O.; Ishitani, R.; Yaremchuk, A.;
Tukalo, M.; Cusack, S.; Sakamoto, K.; Yokoyama, S.
Nat. Struct. Biol. 2003, 10, 425.
17. Nilsson, B.; Moks, T.; Jansson, B.; Abrahmsen, L.;
Elmblad, A.; Holmgren, E.; Henrichson, C.; Jones, T.
A.; Uhlen, M. Protein Eng. 1987, 1, 107.
18. MALDI-TOF MS experiments were conducted on
Applied Biosystems DE-STR mass spectrometer at the
Scripps Center for Mass Spectrometry. The Z-domain
protein solution (0.5 lL, ꢁ0.05 mg/mL) was mixed with
the matrix (0.5 lL, prepared by dissolving a-cyano-4-
hydroxycinnamic acid in a mixture of water and acetoni-
trile or methanol) and deposited on a metal plate that is
used in the mass spectrometer. Data were acquired in
reflector/negative ion mode by using an accelerating
voltage of 25,000 V, grid voltage 90%, guide wire 0.3%,
and an extraction delay time of 250 ns.
24. For the formation of a stable enamine adduct derived
from the b-diketone and the amino group of a lysine
residue activated by burial in a hydrophobic pocket, see:
Barbas, C. F.; Heine, A.; Zhong, G. F.; Hoffmann, T.;
Gramatikova, S.; Bjornestedt, R.; List, B.; Anderson, J.;
Stura, E. A.; Wilson, I. A.; Lerner, R. A. Science 1997,
278, 2085.
25. The imine and enamine adducts derived by conjugation of
an aryl monoketone and the b-diketone to butyl amine,
respectively, were treated with PBS buffer with constant
stirring at pH values of 4.0–10.4 for up to one week and
the percentage of unhydrolyzed imine and enamine
adducts was then assayed using LC/MS. It was found
that the enamine adduct essentially remains intact at
physiological pH and above. In sharp contrast, the imine
adduct is completely hydrolyzed at pH 4.0–10.4 after
overnight stirring.