A highly active DNA polymerase with a fluorous core

Article abstract of DOI:10.1002/anie.200905978

(Chemical Equation Presented) A multifluorinated DNA polymerase with all 14 methionine residues (Met; red in the structure) globally replaced by the non-natural amino acid trifluoromethionine (TFM) exhibits enzymatic activity and selectivity similar to the wild type. The fluorinated enzyme serves as a ¹⁹F NMR probe, and despite its size of 63 kDa, individual ¹⁹F resonances allow the study of enzyme dynamics during DNA synthesis by NMR spectroscopy.

Full text of DOI:10.1002/anie.200905978

Communications
DOI: 10.1002/anie.200905978
Non-natural Amino Acids
A Highly Active DNA Polymerase with a Fluorous Core**
Bastian Holzberger, Marina Rubini, Heiko M. Mꢀller, and Andreas Marx*
DNA polymerases catalyze all DNA synthesis in the cell[1]
and are key tools in important molecular biological core
technologies.[2] Apart from naturally available DNA poly-
merases, several modified DNA polymerases with new
characteristics have been developed. To date, directed
evolution using the 20 natural amino acids is a promising
method for the creation of nucleic acid polymerases with
modified properties.[3] Yet, the incorporation of non-natural
amino acids may lead to enhanced chemical and biological
diversity of protein structures and properties by introduction
of functional groups that are not represented by the natural
amino acids.[4]
DNA polymerase I from Thermus aquaticus (KlenTaq) is a
thermophilic DNA polymerase composed of 540 amino acids
(63 kDa), including 13 methionine (Met) residues[9] that were
globally replaced by TFM with a substitution level of
approximately 82% (Scheme 1). The multifluorinated Klen-
The use of auxotrophic strains, which lack the ability to
biosynthesize one specific natural amino acid, offers the
possibility to replace one amino acid by a non-natural
analogue. After depletion of the natural amino acid in a
defined medium, inducing the expression of the target protein
with concomitant addition of the non-natural analogue leads
to the incorporation of the non-natural amino acid by
selective pressure incorporation.[5]
Fluorinated-protein synthesis has been extensively stud-
ied in the past.[6] It has also been shown that trifluorinated
amino acid analogues such as trifluoromethionine (TFM)
with drastically enhanced hydrophobicity[7] may lead to
proteins with novel characteristics.[4] However, efforts to
engineer new enzyme functions and properties by use of TFM
instead of methionine are hampered, as the replacement of
natural amino acids by trifluorinated analogues can lead to
unfavorable interactions. Thus, the global substitution of
natural amino acids by trifluorinated analogues may cause the
naturally evolved protein scaffold to lose correct folding,
sufficient stability, or enzymatic activity.[4a] Hence, only a few
examples of proteins are known in which methionine was
globally replaced by TFM,[8] including one single example of
an enzyme (phage lysozyme) that is 18 kDa in size, in which
about 70% of the three methionine residues are replaced by
TFM.[8b]
Scheme 1. Multi-TFM-labeled KlenTaq DNA polymerase. a) Methionine
residues (Met) in KlenTaq DNA polymerase (PDB code: 3 KTQ) are
highlighted in red. b) SDS-PAGE gel of purified KlenTaq wt (Met) and
KlenTaq TFM (TFM). M: Marker [kDa].
Taq was highly active and exhibited a similar selectivity as the
wild-type (wt) enzyme. Moreover, the introduction of the
NMR-active nucleus 19F offers the possibility to study DNA
polymerase dynamics by 19F NMR spectroscopy. Despite its
large size of 63 kDa, at least nine individual 19F resonances are
observed, which allow us to distinguish different states of the
DNA polymerase on the way to incorporating a canonical or a
noncanonical nucleotide. To our knowledge, this is by far the
largest enzymatically active protein with Met globally
replaced by TFM.
First, we incorporated TFM into the KlenTaq DNA
polymerase using the Met-auxotrophic E. coli strain B834-
(DE3) in a defined minimal medium. Thus, after depletion of
Met, we incubated the cells for several hours in stationary
phase, washed and resuspended them, added d,l-TFM, and
induced protein expression. We used the racemic mixture of
TFM, as it is assumed that E. coli incorporates only the l-
form into proteins. KlenTaq wt was expressed using the same
E. coli strain in standard LB medium. Both enzymes were
purified to homogeneity (Scheme 1b). The substitution level
of KlenTaq TFM was analyzed by electrospray-ionization
mass spectrometry (ESI-MS; Figure 1 and the Supporting
Information). Comparison of the molecular weights of
KlenTaq TFM and the wild type shows that on average
approximately 82% of the 14 Met amino acids (13 from
KlenTaq gene, 1 initial Met) were replaced by TFM. By
Herein, we present the generation of a multifluorinated
DNA polymerase. The N-terminally truncated version of
[*] B. Holzberger, Dr. M. Rubini, Prof. Dr. H. M. Mꢀller, Prof. Dr. A. Marx
Department of Chemistry and Konstanz Research School Chemical
Biology, University of Konstanz
Universitꢁtsstrasse 10, 78457 Konstanz (Germany)
Fax: (+49)7531-88-5140
E-mail: andreas.marx@uni-konstanz.de
Homepage: www.chemie.uni-konstanz.de/~agmarx/
[**] We gratefully acknowledge Dr. A. Marquardt (Proteomics Facility,
University of Konstanz) for the support with the MS analysis and
financial support by the Konstanz Research School Chemical
Biology and the DFG.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.200905978.
1324
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2010, 49, 1324 –1327

Angewandte
Chemie
To investigate whether the high Met!TFM substitution
level affects the stability of the protein, we studied the
thermostability of KlenTaq TFM in comparison to the wt. The
enzymes were incubated for different periods of time at 958C.
Subsequently, their activity was examined in primer extension
reactions. We observed that KlenTaq TFM lost stability and
exhibited less than 50% primer extension activity after
incubation at 958C for more than 30 min, whereas even
after five hours at 958C the wt showed an activity of 90% (see
the Supporting Information). Nevertheless, it turned out that
KlenTaq TFM is still PCR-active and is still able to amplify
even rather long PCR products (Figure 2b). Furthermore, in
real-time PCR experiments the activity of KlenTaq TFM was
similar to the wt activity (see the Supporting Information).
To quantify the selectivity of KlenTaq TFM, we deter-
mined the error spectra of the enzyme employing a reported
PCR-based assay (see the Supporting Information).[11] The
error rate of 3.2 ꢀ 10ꢀ5 is quite similar to that found for wild-
type KlenTaq (8.8 ꢀ 10ꢀ5).[12] In summary, our studies show
that the global substitution of Met by TFM had most impact
on the stability of the enzyme. Although the CH3!CF3
substitution definitely increases the steric bulk of Met,[4,7]
and despite the fact that the Met residues are mainly located
in the core of KlenTaq DNA polymerase, activity and fidelity
are almost unchanged. Thus, the use of a highly stable protein
enables the incorporation of trifluorinated amino acids with
high substitution levels even in the core of
Figure 1. Deconvoluted ESI mass spectra profiles of KlenTaq wild type
(62883.0 Da, calcd: 62830 Da) and KlenTaq TFM (63504.4 Da). M:
molecular mass. TFM: trifluoromethionine. See also the Supporting
Information.
performing a tryptic digest and analyzing the fragments with a
combination of liquid chromatography (LC) and ESI-MS, we
were able to identify seven of the 14 Met positions (see the
Supporting Information). At position M673 we detected only
TFM, while at the other six positions we found the Met-
bearing peptide fragment as well as the TFM-modified one.
This result indicates that every Met is fluorinated to some
extent.
Next, we investigated whether the highly modified
KlenTaq TFM is still enzymatically active. We carried out
primer extension reactions at temperatures between 37 and
708C (Figure 2a). Both enzymes, KlenTaq TFM and wild
an enzyme without disturbing enzymatic
activity significantly.
Several structural and functional stud-
ies of KlenTaq indicate that the enzyme
undergoes significant conformational
changes upon DNA and nucleotide bind-
ing.[13] The introduction of 19F by TFM
substitution of methionine allows the
dynamics of the enzyme to be studied
during substrate recognition and DNA
synthesis. Thus, we investigated the
enzyme by 19F NMR spectroscopy. Free
KlenTaq TFM without any substrates dis-
plays resonances grouped around the 19F
Figure 2. Enzymatic activity of KlenTaq TFM. a) Primer extension reactions catalyzed by
KlenTaq wt and KlenTaq TFM. Partial primer template sequences employed are depicted on
top. P: primer only; temperature: primer extension at 37.0, 37.8, 40.1, 43.7, 47.6, 51.5, 55.4,
59.4, 63.3, 66.9, 69.2, and 70.08C. b) PCR amplification of a 1655 bp fragment by KlenTaq wt
(Met) and KlenTaq TFM (TFM). PCR program: initial 30 s at 958C and 35 cycles with 20 s at
958C, 60 s at 658C, and 120 s at 728C. M: Marker [bp], wt: wild type.
chemical shift of the TFM amino acid at
d = ꢀ41 ppm. From the expected 14 reso-
nances corresponding to the 14 sequence
positions of TFM, at least nine distinct
resonances (at d = ꢀ38.31, ꢀ39.20, ꢀ39.42,
ꢀ39.68, ꢀ40.12, ꢀ40.52, ꢀ40.95, ꢀ41.03,
and ꢀ41.39 ppm) can be distinguished
type, were able to extend the primer strand to full length in
every case (30 nucleotides (nt)). The incorporation of an
additional nucleotide in a nontemplated manner led to 31 nt
long products. This result has been observed before using 3’!
5’ exonuclease-deficient DNA polymerases.[10] By performing
primer extension reactions at 728C, we determined the
specific activities of the DNA polymerases and observed
that KlenTaq TFM retains more than one-third of the activity
of the wt enzyme (see the Supporting Information). Hence,
the multifluorinated KlenTaq TFM is still a highly active
DNA polymerase, even at temperatures above 708C.
(Figure 3A). These resonances show significant differential
line broadening ranging from 12 Hz of the sharpest signal at
d = ꢀ40.95 ppm to approximately 80 Hz of the broad signals
at d = ꢀ38.31 and ꢀ41.39 ppm.
These differences are most likely caused by different
relaxation times of TFM residues at different positions in the
3D structure of KlenTaq TFM. It can be expected that CF3
groups located at surface-exposed, highly flexible positions
exhibit much smaller line widths than residues in the hydro-
phobic core of the protein, owing to shorter local rotational
correlation times and the lower number of protons as sources
Angew. Chem. Int. Ed. 2010, 49, 1324 –1327
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
1325

Communications
the presence of structural heterogeneity in the
mismatch case at the relatively high concen-
tration of ddCTP used herein to ensure
saturation of the nucleoside triphosphate
binding site. These structural changes might
contribute to the selectivity of KlenTaq TFM
by reducing the efficiency of noncanonical
nucleotide
incorporation.
Thus,
using
19F NMR spectroscopy we were able to clearly
distinguish four different states of the DNA
polymerase on the way to nucleotide incorpo-
ration.
In summary, we report the generation of a
multifluorinated DNA polymerase that exhib-
its nearly the same enzymatic activity and
fidelity as the parental enzyme. Owing to the
high Met!TFM substitution level and the
high signal intensity of the CF3 resonance even
when incorporated into this comparatively
large protein, the KlenTaq TFM serves as an
optimal 19F NMR probe to elucidate altera-
tions of conformation and dynamics during
enzyme catalysis. These findings open up new
Figure 3. 19F NMR spectra of KlenTaq TFM A) without nucleic acid, B) with 2 equiv DNA
substrate (GG), C) with 2 equiv DNA substrate (GG) and after addition of 10 equiv
ddCTP, and D) with 2 equiv of a different DNA substrate (GA) and after addition of over
100 equiv ddCTP. Enzyme concentrations for (A–C) were 42 mm and 18 mm for (D);
intensity was adjusted to (A–C).
of fast dipolar relaxation. Heterogeneity caused by the
incomplete substitution of Met by TFM may broaden the
peaks as well. We therefore conclude that the sharp resonance
at d = ꢀ40.95 ppm, observed in all spectra irrespective of the
presence of substrates, originates from the initial TFM residue
located in the presumably highly flexible N terminus. All
other resonances are more or less strongly affected by adding
substrates. Upon addition of a DNA primer/template com-
plex leading to the formation of a binary protein–DNA
complex, several sharp resonances (full width at half max-
imum less than 40 Hz) show up at new chemical shifts (d =
ꢀ40.03, ꢀ40.37, and ꢀ41.24 ppm), and the appearance of the
other, broader signals also changes, for example, around d =
ꢀ38.31 or ꢀ39.40 ppm (Figure 3B). Subsequent addition of
2’,3’-dideoxycytidine-5’-triphosphate (ddCTP) should lead to
phosphodiester formation with one ddCTP and formation of
the ternary complex with a second ddCTP bound in the active
site.[13] Indeed, the spectrum of the ternary complex (Fig-
ure 3C) is markedly different from those of both the free
DNA polymerase and the binary complex, now with relatively
sharp signals at d = ꢀ40.02, ꢀ40.23, and ꢀ41.30 ppm. Again,
the broader resonances also change their appearance, for
example, around d = ꢀ38.69 or ꢀ39.59 ppm. Taken together,
this evidence is indicative of another conformational change
of KlenTaq TFM upon interaction with ddCTP.
Performing the same experiments with a DNA template
that contains adenine (A) instead of guanine (G) at the
appropriate site in the template leads to the formation of a
mismatched C–A pair upon addition of ddCTP. We found that
the 19F NMR spectrum differs from the match case. Almost all
CF3 resonances of the mismatch complex (Figure 3D) suffer
from severe line broadening. Apart from the signal at d =
ꢀ40.95 ppm that we attribute to the N-terminal TFM residue
(see above), there are no resonances with sharp lines. Instead,
a rather broad signal hump is present around d = ꢀ40.2 ppm.
This finding suggests major changes of protein dynamics or
opportunities for further investigations, such as the directed
evolution of DNA polymerases with new characteristics for
future applications owing to the expanded amino acid
repertoire[3,14] or comparative studies of DNA polymerase
dynamics by 19F NMR spectroscopy, to gain further insights
into DNA synthesis fidelity and substrate spectra.
Received: October 23, 2009
Published online: January 13, 2010
Keywords: DNA polymerases · NMR spectroscopy ·
.
non-natural amino acids · protein engineering ·
trifluoromethionine
[1] A. Kornberg, T. A. Baker, DNA Replication, 2nd ed., Freeman,
New York, 1991.
[2] J. Sambrook, D. W. Russell, Molecular Cloning: A Laboratory
Manual, Cold Spring Harbor Laboratory Press, Cold Spring
Harbor, 2001.
[3] a) D. Loakes, P. Holliger, Chem. Commun. 2009, 4619 – 4631;
b) A. A. Henry, F. E. Romesberg, Curr. Opin. Biotechnol. 2005,
16, 370 – 377; c) S. Brakmann, Cell. Mol. Life Sci. 2005, 62, 2634 –
2646.
[4] a) N. Budisa, Engineering the Genetic Code, Wiley-VCH,
Weinheim, 2006; b) Y. Ryu, P. G. Schultz, Nat. Methods 2006,
3, 263 – 265; c) L. Wang, P. G. Schultz, Angew. Chem. 2005, 117,
34 – 68; Angew. Chem. Int. Ed. 2005, 44, 34 – 66; d) A. J. Link,
M. L. Mock, D. A. Tirrell, Curr. Opin. Biotechnol. 2003, 14, 603 –
609.
[5] C. Minks, S. Alefelder, L. Moroder, R. Huber, N. Budisa,
Tetrahedron 2000, 56, 9431 – 9442.
[6] See, for example: a) J. C. Jackson, J. T. Hammill, R. A. Mehl, J.
Am. Chem. Soc. 2007, 129, 1160 – 1166; b) J. T. Wong, Proc. Natl.
Acad. Sci. USA 1983, 80, 6303 – 6306; c) E. A. Pratt, C. Ho,
Biochemistry 1975, 14, 3035 – 3040; d) O. M. Rennert, H. S.
Anker, Biochemistry 1963, 2, 471 – 476.
[7] E. N. G. Marsh, Chem. Biol. 2000, 7, R153 – R157.
1326
www.angewandte.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2010, 49, 1324 –1327

Angewandte
Chemie
[8] a) N. Budisa, O. Pipitone, I. Siwanowicz, M. Rubini, P. P. Pal,
T. A. Holak, M. L. Gelmi, Chem. Biodiversity 2004, 1, 1465 –
1475; b) H. Duewel, E. Daub, V. Robinson, J. F. Honek,
Biochemistry 1997, 36, 3404 – 3416.
[13] a) W. J. Allen, P. J. Rothwell, G. Waksman, Protein Sci. 2008, 17,
401 – 408; b) P. J. Rothwell, G. Waksman, J. Biol. Chem. 2007,
282, 28884 – 28892; c) P. J. Rothwell, V. Mitaksov, G. Waksman,
Mol. Cell 2005, 19, 345 – 355.
[9] W. M. Barnes, Gene 1992, 112, 29 – 35.
[14] a) C. C. Liu, A. V. Mack, M.-L. Tsao, J. H. Mills, H. S. Lee, H.
Choe, M. Farzan, P. G. Schultz, V. V. Smider, Proc. Natl. Acad.
Sci. USA 2008, 105, 17688 – 17693; b) T. H. Yoo, A. J. Link, D. A.
Tirrell, Proc. Natl. Acad. Sci. USA 2007, 104, 13887 – 13890;
c) J. K. Montclare, D. A. Tirrell, Angew. Chem. 2006, 118, 4630 –
4633; Angew. Chem. Int. Ed. 2006, 45, 4518 – 4521.
[10] J. M. Clark, Nucleic Acids Res. 1988, 16, 9677 – 9686.
[11] P. H. Patel, H. Kawate, E. Adman, M. Ashbach, L. A. Loeb, J.
Biol. Chem. 2001, 276, 5044 – 5051.
[12] C. Gloeckner, K. B. M. Sauter, A. Marx, Angew. Chem. 2007,
119, 3175 – 3178; Angew. Chem. Int. Ed. 2007, 46, 3115 – 3117.
Angew. Chem. Int. Ed. 2010, 49, 1324 –1327
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
1327