Analysis of active site residues in Escherichia coli chorismate mutase by site-directed mutagenesis

Full text of DOI:10.1021/ja953151o

J. Am. Chem. Soc. 1996, 118, 1789-1790
1789
Analysis of Active Site Residues in Escherichia coli
Chorismate Mutase by Site-Directed Mutagenesis
David R. Liu, Sharon T. Cload, Richard M. Pastor, and
Peter G. Schultz*
Howard Hughes Medical Institute, Department of Chemistry
UniVersity of California, Berkeley, California 94720
ReceiVed September 13, 1995
The X-ray crystal structures of three proteins that catalyze
the Claisen rearrangement of chorismate (1) to prephenate (2)
have been solved as complexes with the endo-oxabicyclic
transition state analogue 3 (Scheme 1). Analysis of the
structures of the chorismate mutase from Bacillus subtilis,1,2
the N-terminal 109 amino acid catalytic fragment of the
bifunctional Escherichia coli chorismate mutase-prephenate
dehydratase (“P protein”),3,4 and the catalytic antibody 1F75
indicate that the active sites of both enzyme and antibody
mutases are complementary to the conformationally restricted
transition state analogue 3. In the case of the enzymes, the
structures also reveal a number of groups that could function
in the formation or stabilization of a polar transition state
generated by the heterolysis of the O7-C5 bond.1-5 These
structures, taken together with earlier studies6-13 and recent
mutational analyses of active site residues,14 provide a unique
opportunity to identify common mechanistic features associated
with this novel biological transformation. We describe here
the generation and characterization of 13 active site mutants of
the E. coli monofunctional mutase (EcCM, Figure 1) and
compare the properties of these mutants with the analogous
mutants of the B. subtilis enzyme (BsCM).
In order to facilitate isolation of the wild-type (wt) and mutant
enzymes, six histidines were added directly to the carboxy
terminus of the protein by PCR amplification15 of the EcCM
gene encoded on the plasmid pJS42.16 The resulting histidine-
tagged enzyme was cloned into pAED4,17 an IPTG-inducible
T7 expression vector, to afford plasmid pDRL1. The specific
activity of the histidine-tagged EcCM was comparable to that
of the nontagged mutase as determined by in Vitro expression
and quantitation using [35S]methionine. A series of active site
Figure 1. Active site of EcCM bound to transition state analogue 3.
Scheme 1
mutants (Table 1) was generated using the method of Kunkel,18
and the resulting proteins were expressed in E. coli strain BL2119
and purified on nickel-chelating resin20 (Novagen). The yields
of purified mutases were highly dependent on the nature of the
mutation, ranging from 25 µg/L of culture for Arg28Ala to 6
mg/L for Lys39Ala. Circular dichroism (CD) was used to assess
the secondary structure of mutant proteins (except for Glu52Gln,
Arg28Ala, Arg28Lys, which were not sufficiently pure). The
CD spectra21 of all mutants were superimposable with the wt
spectrum, exhibiting a strong minimum at 220 nm with much
smaller minima at 208 and 212 nm. A small shoulder at ∼210
nm was observed in the CD spectra of the Lys 39 mutants,
raising the possibility that, while still largely helical, these three
mutants may deviate from the precise structure of wt EcCM.
Enzyme activity assays were conducted spectrophotometrically22
by monitoring the conversion of chorismate to prephenate at
pH 7.5 (Table 1).23
The degree to which hydrogen bonding and electrostatic
interactions between active site residues and the C10 and C11
carboxylates of chorismate contribute to catalysis is evidenced
by an analysis of the Arg 11 and Arg 28 mutants. The role of
Arg 11 in hydrogen bonding to the C11 carboxylate is especially
important as indicated by the 103-fold and 104-fold decrease in
kcat/Km for the Arg11Lys and Arg11Ala mutants, respectively,
compared with wild-type mutase. These effects are similar to
those seen for mutants of the corresponding active site residue,
Arg 7, in the B. subtilis enzyme14 and underscore the importance
of the bidentate C11 carboxylate-arginine side chain interaction
in orienting the substrate in the energetically less favored but
reactive pseudodiaxial conformation. In addition to an orien-
tational role, Arg 11 may also exert an electronic effect on the
reaction. Electron-withdrawing groups at C2 of allyl vinyl ether,
corresponding to C8 of chorismate, have been shown experi-
mentally24 and in theoretical models25 to accelerate the nonen-
(1) Chook, Y.-M.; Ke, H.; Lipscomb, W. H. Proc. Natl. Acad. Sci. U.S.A.
1993, 90, 8600.
(2) Chook, Y.-M.; Gray, J. V.; Ke, H.; Lipscomb, W. N. J. Mol. Biol.
1994, 240, 476.
(3) Lee, A. Y.; Karplus, P. A.; Ganem, B.; Clardy, J. J. Am. Chem. Soc.
1995, 117, 3627.
(4) Lee, A. Y.; Stewart, J. D.; Clardy, J.; Ganem, B. Chem. Biol. 1995,
2, 195.
(5) Haynes, M. R.; Stura, E. A.; Hilvert, D.; Wilson, I. A. Science 1994,
263, 646.
(6) Gorisch, H. Biochemistry 1978, 17, 3700.
(7) Addadi, L.; Jaffe, E. K.; Knowles, J. R. Biochemistry 1983, 22, 4494.
(8) Copley, S. D.; Knowles, J. R. J. Am. Chem. Soc. 1987, 109, 5008.
(9) Guilford, W. J.; Copley, S. D.; Knowles, J. R. J. Am. Chem. Soc.
1987, 109, 5013.
(10) Bartlett, P. A.; Nakagawa, Y.; Johnson, C. R.; Riech, S. H.; Luis,
A. J. Org. Chem. 1988, 53, 3195.
(11) Delany, J. J.; Padykula, R. E.; Berchtold, G. A. J. Am. Chem. Soc.
1992, 114, 1394.
(12) Gray, J. V.; Eren, D.; Knowles, J. R. Biochemistry 1990, 29, 8872.
(13) Gray, J. V.; Knowles, J. R. Biochemistry 1994, 33, 9953.
(14) Cload, S. T.; Liu, D. R.; Pastor, R. M.; Schultz, P. G. J. Am. Chem.
Soc. 1996, 118, 1787.
(18) Kunkel Methods Enzymol. 1987, 154, 367.
(19) Grodberg, J.; Dunn, J. J. J. Bacteriol. 1988, 170, 1245.
(20) Hochuli, E.; Dobeli, H.; Schracher, A. J. Chromatogr. 1987, 411,
177.
(15) PCR was executed with the primers below (start and stop codons
are boldfaced, histidine tag is underlined) using a thermal cycle of 94 °C,
30 s; 55 °C, 30 s; 74 °C, 1 min; 25 cycles; 1.5 mM MgCl2): 5′
CTTGACCTGCATATGACATCGGAAAACCCGTTA 3′ (introducing a
NdeI site upstream of the start codon); 5′ CAGGTCAAGAAGCTTT-
TAGTGGTGGTGGTGGTGGTGGAGGATCGATCCGAGAAA 3′ (intro-
ducing a HindIII site downstream of the stop codon).
(16) Stewart, J.; Wilson, D. B.; Ganem, B. J. Am. Chem. Soc. 1990,
112, 4582.
(21) CD spectra were taken at 25 °C in 10 mM Tris pH 7.8, 10% glycerol,
100 mM NaCl at a protein concentration of 30 µg/mL.
(22) The assays were conducted as described in the preceding paper.
(23) Impurities in the preparation were isolated by nickel affinity
chromatography on the cell lysate of untransformed BL21 and possessed
no mutase activity.
(24) Burrows, C. J.; Carpenter, B. K. J. Am. Chem. Soc. 1981, 103, 6983.
(25) Burrows, C. J.; Carpenter, B. K. J. Am. Chem. Soc. 1981, 103, 6984.
(17) Doering, D. S. Thesis, Massachusetts Institute of Technology, 1992.
0002-7863/96/1518-1789$12.00/0 © 1996 American Chemical Society

1790 J. Am. Chem. Soc., Vol. 118, No. 7, 1996
Table 1. Kinetic Constants of EcCM Mutantsa
Communications to the Editor
The side chain of Glu 52 is in close proximity to the C4
hydroxyl of 3 and could potentially act as a general acid or
general base catalyst9 or stabilize charge developing near C4
in a polar transition state. The values of kcat for the Glu52Gln,
Glu52Asp, and Glu52Ala mutants are, respectively, 3.0-, 23-,
and 150-fold lower than that of wild-type mutase. The high
activity of the Glu52Gln mutant argues against a role in which
the carboxylate group of Glu 52 stabilizes developing positive
charge on C4 as well as against mechanisms invoking Glu 52
as a general acid or base. In contrast to the other active site
residues, the activities of mutants of Glu 52 do not parallel those
of the corresponding residue (Glu 78) in the B. subtilis enzyme.14
In the latter case the order of activity is Glu 78 > Asp 78 .
Gln 78 g Ala 78, and thus a carboxylate in the vicinity of the
C4 hydroxyl is likely to be important for catalysis. One possible
explanation for these differences may be that the carboxylate
group of Glu 78 in BsCM is essential for positioning of both
the C4 hydroxyl and the side chain of the critical residue Arg
90.14 In the case of the E. coli enzyme, the Glu 52 side chain
makes no obvious interactions with other critical active site
residues. In addition, modeling suggests that the loss of the
Glu 52-C4 hydroxyl hydrogen bond may lead to an inverted
and nonproductive binding mode26 of chorismate, resulting in
an apparent decrease in kcat. Thus the role of Glu 52 may be
largely orientational in EcCM.
The mutagenesis data of the chorismate mutases from E. coli
and B. subtilis argue against general acid or base catalysis in
the former enzyme and against nucleophilic catalysis in the
latter. These results instead point to the common importance
of hydrogen bonds between mutases and the enolpyruvyl oxygen
and C11 carboxylate of chorismate as most critical to the
catalytic efficiencies of these enzymes. Hydrogen bonds made
to the C4 hydroxyl likely play an orientational role in addition
to the possibility of playing a more complex role in BsCM.
Other residues appear to contribute equally to the binding of
the substrate and transition state. The ability to efficiently
express both enzymes in Vitro will allow us to further probe
the mechanisms of these novel enzymes using unnatural amino
acid mutagenesis.27-29 For example, the electronic versus
orientational effects of the interactions made to O7 are being
examined by replacing Lys 39 and Gln 88 in EcCM with amino
acids which have varying abilities to serve as hydrogen bond
donors such as 5-fluorolysine, 5,5-difluorolysine, 4-fluoro-
glutamine, and thioglutamine.
k
cat/Km
mutant
Km (µM)
kcat (s-1
72b
)
(M-1 s-1
)
Ki (µM)
wt
296 ( 19
>2000
>2000
>2000
>2000
>2000
>2000
>2000
240000
26 ( 4.1
3.66 ( 0.17
R11A
R11K
R28A
R28K
K39A
K39Q
K39R
E52A
E52D
E52Q
Q88A
Q88E
230 ( 10
170 ( 12
230 ( 6.4
4.3 ( 0.57
7.3 ( 0.44
1.9 ( 0.26
4580 ( 450 0.49 ( 0.032 110
218 ( 23
1440 ( 160 3.1 ( 0.20
1080 ( 87 24 ( 1.0
>2000
2200
23000
78.4c ( 4.7
26.8 ( 1.7
12 ( 0.74
361 ( 7.4
300000
210000
12 ( 2.0
>2000
Q88E (pH 4.9) 141 ( 31
wt (pH 4.9)
Q88K
43 ( 4.5
31 ( 2.7
7.3 ( 0.58
147 ( 26
>2000
a The concentration of chorismate used in the assays was limited by
the absorbance range of the spectrophotometer; as a result Km and kcat
for mutants with a Km higher than roughly 2-4 mM could not be
determined. Concentrations of mutant proteins were determined by
comparison with known quantities of wild-type EcCM (determined by
its activity assuming a kcat of 72 s-1)4 on an SDS-PAGE gel stained
with Commassie Blue. Ki values for 3 were determined by monitoring
the reaction at a variety of inhibitor concentrations.15 b Reference 4.
c Velocity vs [3] curve for E52D did not fit standard competitive
inhibition equations at concentrations of 3 g100 µM. At low
concentrations of 3, Ki was determined to be 78 µM.
zymatic Claisen rearrangement of allyl vinyl ether; strong
hydrogen bonds made to C11 in chorismate may promote the
rearrangement in a similar manner. The results of the Arg28Ala
and Arg28Lys mutants, both exhibiting kcat/Km values ap-
proximately 103 times lower than wild-type, also imply a
significant role played by hydrogen bonds made to the C10
carboxylate. This interaction in the E. coli enzyme may be more
important than that in the B. subtilis enzyme, in which the
analogous Arg116Lys mutant exhibits only a 74-fold decrease
in kcat/Km and relatively little change in kcat. These results are
consistent with differences in the three-dimensional structures
of the two enzymes;1,3 in the case of BsCM the C10 carboxylate
of 3 is largely solvent exposed, whereas the transition state
analogue is buried within the EcCM active site.
The ꢀ-amino group of Lys 39 is within hydrogen-bonding
distance of both the ether oxygen (O7) of 3 and an oxygen of
the C11 carboxylate group. The extremely low activities of
the Lys39Ala, Lys39Gln, and Lys39Arg mutants ((3.3 × 104)-
fold to (1.3 × 105)-fold drop in kcat/Km) clearly identify Lys 39
as a critical active site residue. Although the CD spectra of
the Lys 39 mutants differ slightly from that of the wt protein,
it is likely that this residue plays a major catalytic role in the
enzyme. Mutations of the corresponding residue Arg 90 in
BsCM, which also bridges O7 and the C11 carboxylate, result
in similar decreases in kcat/Km.14 The importance of interactions
between EcCM and O7 is further reinforced by the analysis of
mutations to Gln 88, which is also proposed to hydrogen bond
to O7. The Gln88Ala and Gln88Lys mutants manifest a (2 ×
104)-fold decrease in the value of kcat/Km. Given the large effects
of mutations in Lys 39, Arg 90, and Gln 88, it is tempting to
speculate that these residues, in addition to providing an
orientational role in catalysis, stabilize developing negative
charge on the enolpyruvyl moiety in a polar transition state.
Interestingly, the activity of the Gln88Glu mutant, 700-fold
lower than wt at pH 7.5, is rescued almost 103-fold by lowering
the pH to 4.9. This dramatic increase in activity presumably
results from the protonation of Glu 88, reestablishing the
hydrogen bond to O7, and provides convincing evidence that
hydrogen bonding to the enolpyruvyl oxygen of chorismate is
critical to both the kcat and the Km of mutase.
Acknowledgment. This work was supported by a grant from the
National Institutes of Health (R01GM49220). P.G.S. is a Howard
Hughes Medical Institute Investigator. D.R.L. is supported by a
Howard Hughes Medical Institute predoctoral fellowship (HHMI-M940)
and formerly was a National Defense Science and Engineering
Predoctoral fellow (DAAH04-94-G-0296). S.T.C. is supported by a
NIH postdoctoral fellowship (GM15958). We are grateful to Professor
Bruce Ganem for pJS42, to Professor Jon Clardy for the coordinates
of EcCM, and to Professor Jon Clardy and Dr. Sheng Zhang for helpful
discussions. Transition state analogue 3 was generously provided by
Whitney Smith.
JA953151O
(26) Inverting the bound molecule of 3 in the active site such that the
C10 and C11 carboxylates exchange places results in the loss of the two
hydrogen bonds to O7 (and presumably the loss of turnover), but also in
the possible formation of two new hydrogen bonds: the C4 hydroxyl may
hydrogen bond with the side chain N of Gln 88 and with the hydroxyl of
Ser 84. The C11 and C10 carboxylates exchange hydrogen-bonding partners
in this orientation but maintain the number and geometry of these hydrogen
bonds. Perhaps in the Glu52Asp and Glu52Ala mutants, the loss of the
Glu 52-C4 hydroxyl hydrogen bond makes the inverted binding mode more
attractive relative to the normal binding mode, leading to increased
nonproductive binding.
(27) Noren, C. J.; Anthony-Cahill, S. J.; Griffith, M. C.; Schultz, P. G.
Science 1989, 244, 182.
(28) Cornish, V. W.; Mendel, D.; Schultz, P. G. Angew. Chem., Int. Ed.
Engl. 1995, 34, 621.
(29) Ellman, J.; Mendel, D.; Anthony-Cahill, S.; Noren, C. J.; Schultz,
P. G. Methods Enzymol. 1991, 202, 301.