Báo cáo khoa học: Substrate and inhibitor specificity of Mycobacterium avium dihydrofolate reductase pptx

Substrate and inhibitor specificity of Mycobacterium
avium dihydrofolate reductase
Ronnie A. Bock1,*, Jose L. Soulages2 and William W. Barrow1
ă
1 Department of Veterinary Pathobiology, Center for Veterinary Health Sciences, Oklahoma State University, Stillwater, OK, USA
2 Department of Biochemistry and Molecular Biology, Noble Research Center, Oklahoma State University, Stillwater, OK, USA

Keywords
dihydrofolate reductase; mycobacteria; sitedirected mutagenesis; trimethoprim
Correspondence
W. W. Barrow, Department of Veterinary
Pathobiology, 250 McElroy Hall, Center for
Veterinary Health Sciences, Oklahoma State

University, Stillwater, OK 74078, USA
Fax: +1 405 744 3738
Tel: +1 405 744 1842
E-mail:
Website:
*Present address
Department of Biology, University of
Namibia, Windhoek, Namibia
(Received 24 february 2007, revised 22 April
2007, accepted 30 April 2007)
doi:10.1111/j.1742-4658.2007.05855.x

Dihydrofolate reductase (EC 1.5.1.3) is a key enzyme in the folate biosynthetic pathway. Information regarding key residues in the dihydrofolatebinding site of Mycobacterium avium dihydrofolate reductase is lacking. On
the basis of previous information, Asp31 and Leu32 were selected as residues that are potentially important in interactions with dihydrofolate and
antifolates (e.g. trimethoprim), respectively. Asp31 and Leu32 were modified by site-directed mutagenesis, giving the mutants D31A, D31E, D31Q,
D31N and D31L, and L32A, L32F and L32D. Mutated proteins were
expressed in Escherichia coli BL21(DE3)pLysS and purified using His-Bind
resin; functionality was assessed in comparison with the recombinant wild
type by a standard enzyme assay, and growth complementation and kinetic
parameters were evaluated. All Asp31 substitutions affected enzyme function; D31E, D31Q and D31N reduced activity by 80–90%, and D31A and
D31L by > 90%. All D31 mutants had modified kinetics, ranging from
three-fold (D31N) to 283-fold (D31L) increases in Km for dihydrofolate,
and 12-fold (D31N) to 223 077-fold (D31L) decreases in kcat ⁄ Km. Of the
Leu32 substitutions, only L32D caused reduced enzyme activity (67%) and

kinetic differences from the wild type (seven-fold increase in Km; 21-fold
decrease in kcat ⁄ Km). Only minor variations in the Km for NADPH were
observed for all substitutions. Whereas the L32F mutant retained similar
trimethoprim affinity as the wild type, the L32A mutation resulted in a
12-fold decrease in affinity and the L32D mutation resulted in a seven-fold
increase in affinity for trimethoprim. These findings support the hypotheses
that Asp31 plays a functional role in binding of the substrate and Leu32
plays a functional role in binding of trimethoprim.

Dihydrofolate reductase (DHFR, EC 1.5.1.3) is found
in both prokaryotes and eukaryotes, and is essential in
the folate biosynthetic pathway [1]. DHFR catalyzes

NADPH-dependent reduction of dihydrofolate (FAH2)
to tetrahydrofolate (FAH4) (Fig. 1). Reduction of
FAH2 to FAH4 is a universal requirement for the
maintenance of an intracellular reduced folate pool,
which is important in one-carbon transfer reactions

that are necessary for the biosynthesis of DNA, RNA
and protein [2,3]. A common bacterial antifolate inhibitor of this enzyme is trimethoprim (TMP) (Fig. 2).
Previously, we identified the Mycobacterium avium
folA gene and confirmed the functionality of its product, DHFR [4]. Subsequently, we demonstrated its
inherent resistance to TMP [5,6]. Further studies
revealed a group of 2,4-diamino-5-deazapteridines that

Abbreviations
DHFR, dihydrofolate reductase; FAH2, dihydrofolate; FAH4, tetrahydrofolate; rDHFR, recombinant dihydrofolate reductase; TMP,
trimethoprim.

3286

FEBS Journal 274 (2007) 3286–3298 ª 2007 The Authors Journal compilation ª 2007 FEBS


R. A. Bock et al.
ă

Site-directed mutagenesis of M. avium DHFR

Dihydropteridine
Dihydropteroate synthase [2.5.1.15]

Dihydropteroate
Antifolates:
Methotrexate
Trimethoprim
Deazapteridines

Dihydrofolate synthase [6.3.2.12]

Dihydrofolate
{NADPH + H2Folate

NADP+ + H4Folate}

Dihydrofolate reductase [1.5.1.3]

Tetrahydrofolate

Fig. 1. FAH4 biosynthesis, including

enzymes and EC numbers. Examples of
common antifolates are given above reaction formulae.

Thymidylate
Nucleic Acids

O

COOH

OH
N3

4

N

2 1

N

H2N

N

H

N
H
N

COOH

dihydrofolate

NH2
O

N3
2

H2N

4
1

CH3
O

N

O

Amino Acids

CH3
CH3
Trimethoprim

Fig. 2. Chemical structures of the DHFR substrate FAH2 and the
antifolate TMP.

are effective against M. avium because of their activity

against M. avium DHFR but not human DHFR [6].
Continued efforts in the development of better antifolate derivatives in this class will depend upon a better understanding of the M. avium DHFR active site.
In that regard, we have initiated mutagenesis studies
to evaluate individual amino acids located in key positions of the DHFR binding site.
Although a crystal model of M. avium DHFR has
not been published, one published manuscript is useful
for understanding certain molecular aspects of this
enzyme. Kharkar and Kulkarni developed a homology
model of M. avium DHFR based on the X-ray crystal

structure of M. tuberculosis and homology with the
M. avium DHFR sequence [7]. With the use of that

model, important amino acids were identified by
so-called ‘comparative protein modeling’ or ‘homology
modeling’ with the previously published M. tuberculosis DHFR crystal structure [8] and analysis of published inhibitor data [7]. From Kharkar’s model, it was
determined that M. avium DHFR has the same general
fold that is found in other bacterial DHFRs, namely a
central b-sheet flanked by a-helices [7]. The authors
proposed that structure–function studies using sitedirected mutagenesis would be essential to verify the
importance of specific amino acid residues in the binding site, particularly Asp31 and Leu32 [7].
The negatively charged Asp31, equivalent to the
Asp27 in M. tuberculosis DHFR, is a highly conserved amino acid found in most bacterial DHFRs
[7]. It is located in the substrate-binding site and, on
the basis of other bacterial DHFRs, is most likely

important in catalysis [7]. In Escherichia coli DHFR,
the importance of the equivalent position (Asp27) in
catalysis has also been suggested [9,10]. This was later
confirmed by Howell et al. using mutagenesis [11]. An
identical function for this equivalent residue in Lactobacillus casei DHFR (D26) has also been reported
[12,13].
The negatively charged carboxyl moiety is essential
in the hydrogen bonding that takes place between the
natural substrate’s 2-amino group and the N3 position
in the pteridine ring (Fig. 2). This similar interaction
takes place with inhibitors whose activity is based

FEBS Journal 274 (2007) 3286–3298 ª 2007 The Authors Journal compilation ª 2007 FEBS

3287


Site-directed mutagenesis of M. avium DHFR

R. A. Bock et al.
ă

upon a structural resemblance to FAH2. A list of these
types of inhibitor would consist of TMP, methotrexate,

and 5-deazapteridine derivatives [7], including those
described in our previous studies [6,14]. Our hypothesis
was that modifications of the M. avium DHFR at position D31 would substantially affect substrate binding
and enzyme efficiency.
Leu32 is located in a hydrophobic area of the
M. avium DHFR FAH2-binding site [7]. In this position, Leu32 would presumably interact with the end of
antifolate inhibitors distal to the pteridine ring (e.g.
methotrexate and TMP) [7]. The Leu32 residue is
equivalent to Phe31 and Gln28 in human and
M. tuberculosis DHFRs, respectively [7]. This rationale
is also suggested by equivalent residues in two other
bacterial DHFRs, those of E. coli and L. casei [12].

We hypothesized that modification of Leu32 in M. avium DHFR would affect binding and the IC50 of TMP
but not of the normal substrate. The site-directed
mutants in this study were designed in order to
confirm the functional importance of D31 and L32
as substrate-binding, cofactor-binding and inhibitorbinding sites.
The objectives of this study were to: (a) utilize sitedirected mutagenesis to modify Asp31 and Leu32 residues in the M. avium recombinant DHFR (rDHFR);
(b) assay the enzyme activity and kinetic parameters of
the mutated products; (c) verify the results using complementation with an E. coli DHFR-deficient mutant;
and (d) obtain IC50 values for TMP with mutated
rDHFRs.

Results

Table 1. M. avium folA D31 and L32 mutations and primers used
to construct them.

Mutation

Primers (5¢- to 3¢) with mutational
codon underlined

D31E
D31Q
D31A

D31N
D31L
V76A
L32F
L32A
L32D

CGAGGAGCTCACCCGGTTCAAG
CGTGCCCGAGCAACTCACCCGGTTC
CGAGGCCCTCACCCGGTTCAAAG
GCCCGAGAACCTCACCCGGTTCAAAG
CGTGCCCGAGCTCCTCACCCGGTTCAAAG

CCCGACTTCGCCGCCGAGGGG
GCCCGAGGACTTCACCCGGTTC
GCCCGAGGACGCCACCCGGTTC
GCCCGAGGACGACACCCGGTTC

Fig. 3. SDS ⁄ 12.5% polyacrylamide gel showing steps involved in
the purification of M. avium rDHFR. Lane 1 contains Novagen Perfect protein markers (sizes in kDa are indicated on the left of the
figure). Lane 2 contains precolumn protein extract. Lane 3 contains
flow-through. Lane 4 contains wash with 5 mM imidazole. Lane 5
contains wash with 40 mM imidazole. Lanes 6–13 contain purified
rDHFR that eluted in successive fractions from 150 to 300 mM imidazole. These fractions all showed activity in the standard DHFR
assay described in Experimental procedures. Similar results were

obtained with the mutated rDHFRs.

Mutagenesis
Sequencing data showed that the GeneEditor mutagenesis protocol used to construct the M. avium DHFR
mutants was highly efficient. Efficiencies of 80% and
higher were achieved. The mutagenic oligonucleotides
are listed in Table 1.
Protein expression and purification
His-Bind resin was used to recover recombinant wildtype and mutant M. avium DHFR from the soluble
fraction of a cell extract under nondenaturing conditions. In this semi-automated process, His-Bind resin
columns were loaded and washed manually and by
gravity flow until the eluate showed no significant

reading at 280 nm. The automated elution of the HisBind resin-bound protein through a 5–500 mm imidazole linear gradient is shown in Fig. 3 for the wild-type
3288

rDHFR. The flow-through fraction (Fig. 3, lane 3)
shows that most DHFR was bound to the resin. The
bound protein eluted between 150 and 300 mm imidazole (Fig. 3). The elution profiles for the recombinant
wild-type and mutant DHFR were consistent and similar. The percentage yield of active enzyme was 34%.
This corresponds to an increase in purification of
between 10-fold and 14-fold.
CD
Figure 4 shows the far-UV CD spectra of wild-type
DHFR and some of the mutants used in this study.

Consistent with the a ⁄ b-sheet motif, the spectra of the
proteins show a minimum at 215 nm and become positive below 200 nm. The fractions of four structural
components (a-helix, b-sheet, b-turns and unordered)

FEBS Journal 274 (2007) 3286–3298 ª 2007 The Authors Journal compilation ª 2007 FEBS


R. A. Bock et al.
ă

Site-directed mutagenesis of M. avium DHFR

residues to the peptide CD spectrum [17]. Because the
spectroscopic properties of the multiple Trp residues of
DHFR have also been shown to be sensitive to several
mutations, the spectra and estimates of secondary
structure of the mutants could be affected by the properties of the Trp residues. Despite this consideration,
overall, the CD data suggest that the mutants retained
the a-helix ⁄ b-sheet motif that characterizes DHFR.
Asp31 mutations

Fig. 4. Far-UV CD spectra of DHFR wild type and mutants. Spectra
were acquired in 20 mM sodium phosphate, 100 mM NaCl (pH 7.4)
at 25 °C. With exception of the spectrum for the L32D mutant,

which was obtained from a single protein sample, the spectra
shown represent the average obtained from at least two independent protein preparations. Protein names and corresponding symbols
are indicated.

were estimated by the ‘self-consistent’ method (selcon3) according to Sreerama et al. [15]. These estimates are shown in Table 2. The inferred secondary
structure of wild-type M. avium DHFR is consistent
with the crystal structures of the homologous DHFRs
from M. tuberculosis [8] and E. coli [16], whose structures comprise eight b-strands and four a-helices. The
structural information obtained by CD spectroscopy is
also consistent with the homology model of M. avium
[7]. The deduced secondary structure of the mutants
(Table 2) suggests that the mutations did not promote

significant structural changes in the secondary structure of the protein. Even though the estimated structures of the mutants D31Q and L32D do not differ
significantly from the structure of the wild-type protein, their CD spectra show some differences from the
spectra of the other proteins. We do not know the reason for these spectral differences. Previous studies with
DHFR mutants have shown a contribution of the Trp

Table 2. Secondary structure of DHFR wild type (WT) and mutants.
The fractions of different structural components were calculated
from the CD spectra using the program SELCON3.
a-Helix
WT
D31A
D31E

D31L
D31N
D31Q
L32D

b-Sheet

b-Turn

Unordered

Sum

0.20
0.20
0.20
0.19
0.15
0.21
0.16

0.38
0.38
0.35

0.28
0.30
0.38
0.36

0.17
0.17
0.15
0.22
0.23
0.20
0.17

0.25
0.25
0.22
0.31
0.31
0.20
0.25

1.00
1.00
0.92

1.00
0.99
0.99
0.93

His-Bind resin eluted fractions were assayed for
enzyme activity in an in vitro enzyme assay, and the
results show a significant reduction in enzyme activity
for all the Asp31 mutants assayed, D31A, D31E,
D31Q, D31N and D31L. Table 3 shows that although
these mutant M. avium rDHFRs still displayed functionality, the activity was significantly reduced. The
D31A and D31L mutants showed a reduction in

enzyme activity of over 90%, and the D31E, D31Q
and D31N mutants showed reductions of 81.1%,
85.3% and 84.6%, respectively, as compared to the
wild-type M. avium rDHFR (Table 3). However, the
negative control mutant V76A did not show a reduction in activity as compared to the wild-type rDHFR
(Table 3).
Using the sas statistical software, Dunnett’s test
shows (at significance level a ¼ 0.05) that with the
exception of the control mutant V76A, the enzyme
specific activities of all Asp31 mutants were significantly different from that of the wild-type rDHFR
(p807) (Table 3). The control mutant V76A, which has
a modification outside the enzyme’s active site, still

had a specific activity that amounted to 99% of that
of the wild-type rDHFR and was therefore statistically
not different (Table 3).
Leu32 mutations
In contrast to the Asp31 substitutions, only one of the
Leu32 substitutions caused a significant change in the
enzyme’s specific activity as compared to the wild-type
rDHFR (Table 3). Neither the L32F nor the L32A
substitution affected the enzyme’s interaction with
FAH2. In comparison to the wild-type rDHFR, the
L32F mutant had a slightly higher relative specific
activity (103%), showing that, statistically (Dunnett’s

test P ¼ 0.759 at a ¼ 0.05), this mutant was not different from the wild-type rDHFR (Table 3). The L32A
mutant of the M. avium DHFR still had 96% of the
specific activity of the wild-type rDHFR, and was
therefore statistically not different from the wild-type
rDHFR (Dunnett’s test P ¼ 0.742 at a ¼ 0.05). The

FEBS Journal 274 (2007) 3286–3298 ª 2007 The Authors Journal compilation ª 2007 FEBS

3289


Site-directed mutagenesis of M. avium DHFR

R. A. Bock et al.
ă

Table 3. Comparison of enzyme activity of wild-type rDHFR (p807) with those of the D31, V76 and L32 mutated DHFRs. Assays were performed in triplicate, and results were evaluated statistically using the Dunnett post-test. P-values are given for significance level a ¼ 0.05.

DHFR

Specific activity
(lmolỈmin)1Ỉmg)1)

Percentage relative

specific activity

Percentage decrease
in specific activity
over wild type

P

p807
D31A
D31E
D31Q

D31N
D31L
V76A
L32F
L32A
L32D

15.4
1.05
2.91
2.26
2.37

0.060
15.2
16
14.7
5.09

100
6.82
18.9
14.7
15.4
0.39

98.7
103
96
33

–
93
81
85
85
99.6
1.3

–
4
67

–
0.0001
0.0001
0.0001
0.0001
0.0001
0.914
0.759

0.742
< 0.0001

L32D substitution was the only one in this series that
had a negative impact on the mutant enzyme’s activity.
The L32D mutant had a 67% reduction in its specific
activity as compared to the wild-type rDHFR, and
was therefore statistically different from the wild-type
rDHFR (Dunnett’s test P < 0.0001 at a ¼ 0.05).
Kinetic characteristics of Asp31 mutations
The kinetic parameters Km and Vmax for FAH2 and
NADPH of the wild-type M. avium rDHFR (p807)

and the Asp31 mutants were determined with the
nonlinear Michaelis–Menten curve-fitting program
(Table 4). The D31A and D31L mutants show the
greatest changes in kinetic characteristics as compared
to the wild-type rDHFR (p807). The Km (FAH2) was
37 lm for the D31A mutant and 198 lm for the D31L

<
<
<
<
<

mutant, as compared to 0.70 lm for p807. This corresponds to a 51-fold increase for the D31A mutant and
a 283-fold increase for the D31L mutant. The
Km (FAH2) values for the D31E, D31Q and D31N
mutants were 1.92, 2.32 and 2.08 lm, respectively; a
moderate increase of 2–2.5-fold over p807. The D31E
mutant showed a slight increase in Km (NADPH) over
the wild type. In contrast to the Km (FAH2), which
showed a 283-fold increase, the Km (NADPH) of
1.44 lm for the D31L mutant was not too different
from that of the wild type. The D31A, D31Q and
D31L mutants had the lowest kcat ⁄ Km (FAH2) values:

2.3 · 104, 65 · 104 and 0.013 · 104 m)1Ỉs)1, respectively. The values for the D31E and D31N mutants
were very similar, at 187 · 104 and 240 · 104 m)1Ỉs)1,
respectively (Table 4). The Km (NADPH) for the
D31A mutant of 0.68 lm was similar to the values for

Table 4. Kinetic parameters at pH 7.0 and 30 °C of recombinant wild-type p807 vs. the D31, V76 and L32 mutants of M. avium DHFR for
FAH2 and NADPH, determined with the nonlinear Michaelis–Menten curve-fitting program ENZFITTER. Each mutant was assayed in triplicate
with at least 10 different substrate concentrations. The results are expressed as the mean ± SEM. The molecular mass of DHFR used in
the calculation of the kcat values was 20 000 Da.
NADPH

FAH2

DHFR
p807
D31A
D31E
D31Q
D31N
D31L
V76A
L32F
L32A
L32D

0.7
37
1.92
2.32
2.08
198
0.78
0.68
0.96
5.12

kcat
(s)1)

Km
(lM)

3290

±
±
±
±

±
±
±
±
±
±

0.034
1.68
0.092
0.085
0.100

5.53
0.033
0.047
0.046
0.252

kcat ⁄ Km
(· 104 M)1Ỉs)1)

Km
(lM)

20.5
0.86
3.58
1.5
5.03
0.025
18.7
21
16.2
7.2

2900

2.3
187
65
240
0.013
2400
3090
1690
141

1.55
0.68

2.01
0.65
0.77
1.44
1.56
1.32
1.39
0.84

kcat
(s)1)
±

±
±
±
±
±
±
±
±
±

0.0465
0.0269

0.0953
0.0288
0.0321
0.067
0.065
0.056
0.0551
0.040

kcat ⁄ Km
(· 104 M)1Ỉs)1)

26.8
0.59
2.05
2.18
8.50
0.02
23.7
23.2
17.8
10

1730

38
102
335
1100
1.39
1520
1760
1280
1190

FEBS Journal 274 (2007) 3286–3298 ª 2007 The Authors Journal compilation ª 2007 FEBS

R. A. Bock et al.
ă

Kinetic characteristics of Leu32 mutants
The kinetic parameters Km and Vmax for FAH2 and
NADPH were also determined at pH 7.0 and 30 °C,
and are listed in Table 4 for the wild type and the
Leu32 mutants of M. avium DHFR. The L32F and
L32A mutants were not very different from the wild
type in Km (FAH2) values. The Km (FAH2) of the wild
type was 0.7 lm, whereas that of the L32F mutant was

0.68 lm. The L32A mutant had a slightly higher
Km (FAH2) of 0.96 lm. With a Km (FAH2) of
5.12 lm, the L32D mutant had a seven-fold increase in
Km (FAH2) over the wild type.
The L32F and L32A mutants were also not much
different from the wild type in their Vmax (FAH2) values (not shown). The wild type had a Vmax (FAH2) of
61.1 lmolỈmin)1Ỉmg)1, whereas the L32F and L32A
mutants had a Vmax (FAH2) of 63.3 and 55.2 lmolỈ
min)1Ỉmg)1, respectively. The kcat ⁄ Km (FAH2) of the
L32F mutant was slightly higher than that of the wild
type. The kcat ⁄ Km (FAH2) of the L32F mutant was
3090 · 104 m)1Ỉs)1, whereas that of the wild type was

2900 · 104 m)1Ỉs)1 (Table 4). The L32A mutant had a
lower kcat ⁄ Km (FAH2) as compared to the wild type.
At 1690 · 104 m)1Ỉs)1, this mutant’s kcat ⁄ Km (FAH2)
value was almost 1.5-fold lower than that of the wild
type (Table 4). On the other hand, the L32D mutant
had a kcat ⁄ Km (FAH2) value of 141 · 104 m)1Ỉs)1,
which was almost 21-fold lower than that of the wild
type (Table 4). The wild-type M. avium DHFR and
the L32F and L32A mutants were also similar with
respect to NADPH binding. The wild type had a
Km (NADPH) of 1.55 lm, whereas the Km values of
the L32F and L32A mutants were 1.32 and 1.39 lm,

respectively (Table 4). The Km (NADPH) of the L32D
mutant was 0.84 lm, and therefore almost two-fold
lower than that of the wild type. The wild type and
the L32F and L32A mutants had Vmax (NADPH) values of 80, 69.6 and 53.5 lmolỈmin)1Ỉmg)1, respectively
(Table 4), whereas Vmax (NADPH) of the L32D
mutant was 30.1 lmolỈmin)1Ỉmg)1.
Growth complementation
The DHFR-deficient E. coli strain MG1655folA::kan3
(MH831) [40] was transformed with plasmid pET15b

containing the wild-type M. avium DHFR (MH831 +
p807), the DHFR Asp31 mutant forms (p807D31,

p807D31A, p807D31E, p807D31Q, p807D31N, and
p807D31L), or the mutation control V76A
(p807V76A) and the Leu32 mutant forms (p807L32F,
p807L32A, and p807L32D). In addition, the DHFRdeficient strain was also transformed with pET15b that
did not contain the M. avium DHFR gene insert
(folA–pET15b) as a vector control. MH831 is unable
to grow in the absence of thymidine [40].
All of the substitutions within the plasmid-borne
M. avium DHFR enabled growth of MH831 that was
comparable to that of the parent strain MG1655, in
the presence of thymidine (data not shown). However,
in the absence of thymidine, the DHFR-deficient

MH831 did not grow, and none of the p807 Asp31
DHFR mutants was able to restore growth by complementation (Fig. 5). Only the wild-type M. avium
DHFR (p807) and the control DHFR mutant
p807V76A restored growth of MH831 to levels comparable to that of the E. coli MG1655 parent strain
(Fig. 5). The DHFR-deficient strain that was transformed with pET15b vector control also did not show
growth complementation in the absence of thymidine
(data not shown).
The L32 mutations of the M. avium DHFR were
able to complement the DHFR-deficient E. coli strain
and restore growth to levels comparable to that of its
parent strain, E. coli MG1655, even in the absence of
2.5

MG1655
MH831 + p807
MH831 + p807V76A

2.0

MH831 + p807L32F
MH831 + p807L32A
MH831 + p807L32D

1.5

D 600

the D31Q and D31N mutants. The data indicate that
the control mutation V76A did not affect the Km for
FAH2 or NADPH (Table 4). The V76A mutant had a
Km (FAH2) of 0.78 lm, as compared to 0.7 lm of the
wild type, and a Km (NADPH) of 1.56 lm, as compared to 1.55 lm of the wild type.

Site-directed mutagenesis of M. avium DHFR

1.0

0.5

0.0
0

2

4
Time (h)

6

8

Fig. 5. Example of growth curve of DHFR-deficient E. coli strain
MH831, transformed with wild-type M. avium DHFR p807
(MH831 + p807), various recombinant mutated M. avium DHFRs,
the Leu32 mutants, the control mutant V76A, and the MH831 parent strain MG1655. All strains grew in the presence of thymidine
(data not shown). Only the strains plotted in the figure grew in the
absence of thymidine.

FEBS Journal 274 (2007) 3286–3298 ª 2007 The Authors Journal compilation ª 2007 FEBS

3291

Site-directed mutagenesis of M. avium DHFR

R. A. Bock et al.
ă

Table 5. IC50 of TMP for recombinant wild-type and Leu32 mutants
of M. avium DHFR as determined by the four-parameter curve procedure (Bio-TEK enzyme software). Assays were done in triplicate.
Results are shown as the mean ± SEM.
IC50 (nM)
DHFR

TMP

Wild type
L32F
L32A
L32D

3900
3600
45 000
570

±
±
±
±

425
143
3406
68

thymidine (Fig. 5). However, in the absence of thymidine, a five-fold reduction in growth was observed with

the L32D mutant, and a reduction of growth with the
L32A mutant was also observed to a lesser degree
(Fig. 5).
IC50 assay
The IC50 values of the wild-type rDHFR and the
L32F mutant of M. avium DHFR for TMP were
very similar, at 3900 nm and 3600 nm, respectively
(Table 5). The IC50 value of 45 000 nm for the L32A
mutant for TMP, however, was about 12-fold higher
than that of the wild-type rDHFR (Table 5). On the
other hand, the L32D mutant’s IC50 value for TMP
was 570 nm, which was about seven-fold lower than

that of the wild-type rDHFR (Table 5).

Discussion
Site-directed mutagenesis has previously been used to
investigate the role of the highly conserved active site
carboxylic acid residue in catalysis and binding of
FAH2 in E. coli, L. casei and a few other species
[1,11,18–22]. Various studies have also investigated
hydrophobic interactions within the binding cavity.
This is the first study that has assessed the important
functional role of the highly conserved D31 and conserved L32 residues in M. avium, or equivalent residues
in any other mycobacterial DHFR.

In this study, all mutations of Asp31 (D31) affected
the M. avium DHFR, causing significant reductions in
the enzyme’s interaction with FAH2. The V76A substitution, theorized to be outside of the binding cavity,
did not cause a change in the enzyme’s specific activity
or kinetics as compared to the wild-type rDHFR. This
control validated the mutation procedure.
The D31A substitution removed the side chain and
charge at that position, thus resulting in a reduction of
specific activity by almost 95%. This is consistent with
3292

this residue’s role in other DHFRs described thus far,

and a reflection of its strictly conserved status for all
DHFRs [1,8,11,23–30]. This severe decrease in enzyme
functionality is also reflected in the kinetics. The
kcat ⁄ Km (FAH2) decrease over the wild-type rDHFR
for the D31A mutant was over 1200-fold.
The D31E substitution with the carboxylic acid
group of the wild type caused a reduction in specific
activity of 80%, suggesting that the larger side chain
(i.e. methylene group) has a negative effect on interactions within the binding cavity. Like the M. avium
D31E mutant in this study, David et al. [31] found
that a D27E mutated E. coli DHFR was 17-fold less
efficient than the wild type. The D31N and D31L

mutants both have a side chain size of similar size to
the wild-type rDHFR’s aspartic acid. However,
although the D31N side chain is not charged, it is
nonetheless polar. The D31L side chain is not charged
and is nonpolar. The D31N mutant showed  85%
reduction in specific activity over the wild type, therefore behaving similarly to the D31E mutant. Howell
et al. [11] reported a D26N substitution of E. coli
DHFR that severely affects enzyme function. That
mutated enzyme had < 1% of the specific activity of
the wild type, and severely altered kinetics [11]. Basran
et al. [32] reported that a D26N substitution of
L. casei resulted in a smaller reduction of the enzyme’s

functionality as compared to an equivalent substitution
in E. coli DHFR [32]. The L. casei D26N DHFR had
a decrease in kcat (FAH2) of nine-fold, compared to
300-fold in E. coli, and a decrease in kcat ⁄ Km (FAH2)
of 13-fold, compared to 11 000-fold in E. coli [11,32].
The equivalent substitution in M. avium DHFR
(D31N) has effects that resemble those for L. casei
DHFR more than those for E. coli DHFR. The
M. avium D31N DHFR had a four-fold decrease in
kcat (FAH2) and a 12-fold decrease in kcat ⁄ Km (FAH2).
These studies indicate that the asparagine substitution
in this position affects the enzyme’s activity in M. avium, E. coli and L. casei.

The D31L mutant of M. avium DHFR had < 1%
of the wild-type rDHFR’s activity. The very low specific activity of the D31L mutant was also reflected in
the mutant’s kinetic behavior. David et al. [31] found a
D27L mutant of E. coli DHFR to be similarly dysfunctional. This substitution is similar in side chain
size to the wild-type rDHFR’s aspartic acid, but has
no charge and is nonpolar. It is the most severe change
of all the D31 substitutions, suggesting that the
increased hydrophobicity had more severe negative
effects than the loss of the charge alone.
The D31Q substitution in M. avium DHFR,
which is also one methylene group larger than the

FEBS Journal 274 (2007) 3286–3298 ª 2007 The Authors Journal compilation ª 2007 FEBS


R. A. Bock et al.
ă

residue in the recombinant wild-type enzyme and has
the same size as the D31E substitution, did not have
the charge associated with the carboxylic acid group.
This mutant also showed a reduction in specific activity
of > 80%, or about the same as the reductions seen
for the D31E and D31N mutants. The kcat ⁄ Km (FAH2)

for the D31Q mutant was decreased by 45-fold in comparison to the wild-type rDHFR, as compared to 16fold and 12-fold reductions for the D31E and D31N
mutants, respectively (Table 4). No previous studies
with this type of substitution have been performed.
The Km (NADPH) values of all D31 mutants were
slightly reduced in comparison to those of the wildtype rDHFR. However, this reduction did not vary
greatly among the D31 mutants, for which the values
ranged from 0.65 to 0.77 lm, representing about a
two-fold reduction as compared to the wild type.
Dunn et al. [20] and Appleman et al. [18] found that
replacement of the conserved D27 in E. coli reduced
the affinity of NADPH by seven-fold and three-fold,
respectively. Although this conserved aspartic acid residue is not directly involved in NADPH binding, they

argued that the increased rate of dissociation in the
mutants, as well as a shift in the equilibrium that
favors nonbinding, could be responsible for changes in
NADPH affinity. On the basis of the Km (NADPH)
for all D31 mutants in this study, the effects of the
substitutions on binding of NADPH seem to be minimal in comparison to the effects on binding of FAH2.
From these results, it is evident that replacement of
Asp31, regardless of the substitution, severely affects
M. avium DHFR function. This is further supported
by the fact that none of the D31 mutants was able
to restore growth to the DHFR-deficient E. coli. Only
the wild-type M. avium rDHFR and the recombinant

V76A mutant were able to restore MH831 growth to
levels comparable to that of the MG1655 parent strain.
We can conclude that Asp31 has an important functional role in catalysis in M. avium DHFR, and that it
is unlikely that any substitution at this site would
result in a functional enzyme.
As in all DHFRs [21,23,33–36], the conserved Leu32
is one of several hydrophobic residues that line the
M. avium DHFR binding cavity. Unlike D31, which is
only substituted by glutamic acid in vertebrates, L32 is
also replaced by glutamine in some bacteria, whereas
vertebrates have either a phenylalanine or a tyrosine in
this position [35]. In contrast to the D31 substitutions,

only one of the L substitutions L32D caused a significant change in the enzyme’s normal reaction with
FAH2. These results indicate that L32 is not directly
involved in catalysis; this was also indicated by the
mutant’s kinetic behavior. The kinetics of the L32F and

Site-directed mutagenesis of M. avium DHFR

L32A mutants were similar to those of the wild type for
both FAH2 and NADPH, whereas the L32D mutant
showed a 21-fold decrease in kcat ⁄ Km (FAH2). Equivalent substitutions in E. coli, mouse and human rDHFRs
do not give similar results [37–39]. E. coli DHFR has no
increased catalytic efficiency resulting from an L28F

substitution [38], and no change in kinetics resulting
from an L28Y substitution [37]. Whereas this study did
not find a change in the activity of the L32A mutant as
compared to the wild-type M. avium rDHFR, Chunduru et al. [39] found that an equivalent substitution
(F31A) in human DHFR caused a four-fold higher
Km (FAH2) and a four-fold reduced kcat ⁄ Km. No prior
information regarding an L32D substitution was found
in literature searches. However, the introduction of a
charged group with the aspartic acid substitution may
have adverse effects on interactions in the binding
cavity. Baccanari et al. [9] found two E. coli (RT500)
DHFR isozymes, one of which had L28, and the other

of which had R28. They argued that interaction between
R28 and the conserved D27 probably led to reduced efficiency. It is possible that the same could occur with the
L32D substitution in the M. avium DHFR.
The L32F substitution did not change the affinity for
TMP (Table 5). This is consistent with findings that an
equivalent substitution in human rDHFR does not alter
the affinity for TMP [35]. The L32A substitution
removed the hydrophobic residue from this position,
which may have caused a change in local hydrophobicity. X-ray crystal structures have shown that the equivalent hydrophobic residue in E. coli and L. casei DHFR
is in contact with inhibitors. In their proposed model of
the M. avium DHFR, Kharkar & Kulkarni [7] also
point to the role of L32 interacting with inhibitors. The

current finding supports this conjecture for M. avium
DHFR. The decrease in TMP affinity caused by the
M. avium L32A substitution was also observed with the
equivalent substitution (F31A) in human DHFR [39].
The L32D substitution in M. avium DHFR resulted
in a seven-fold increase in the affinity for TMP
(Table 5). An introduction of the charge associated
with aspartic acid seems to stabilize TMP but not
FAH2. In the X-ray crystal structures of L. casei and
E. coli DHFR, the p-aminobenzoic acid ring of folate
is closely aligned with L27 and L28, respectively
[10,13]. Kharkar et al. [7] proposed the same for L32

in M. avium DHFR. The L32D substitution in M. avium did not only remove the hydrophobic residue that
stabilized FAH2, but also introduced a charge that
may destabilize FAH2 in the binding cavity. Structurally, TMP does not have the hydrophobic ring of
FAH2; it does have methoxy groups that might interact with aspartic acid to stabilize the inhibitor.

FEBS Journal 274 (2007) 3286–3298 ª 2007 The Authors Journal compilation ª 2007 FEBS

3293


Site-directed mutagenesis of M. avium DHFR

R. A. Bock et al.
ă

The results of this study are consistent with what
is currently known about the highly conserved binding cavity aspartic acid residue in E. coli, L. casei
and other DHFRs. It is apparent that Asp31 in
M. avium DHFR plays a functional role in catalysis.
Modifications in size (D31E), size and charge (D31A
and D31Q) and charge (D31N and D31L) result in
a significant reduction of normal enzyme activity,
primarily by affecting the binding of FAH2. In addition, the results of this study also support the hypothesis that L32 plays a more functional role in the
binding of antifolate inhibitors such as TMP, as

opposed to FAH2. It has been shown that certain
modifications of L32 in M. avium DHFR can affect
the affinity of the enzyme for TMP, a drug to which
the organism is naturally resistant. Whereas an L32F
substitution has no effect on affinity, an L32A substitution decreases the affinity for TMP, and an
L32D substitution increases the affinity for the antifolate. Further studies are underway to determine the
effect of these and other mutations on the binding
of other antifolates.

Large-scale protein expression and purification
of wild-type and mutant DHFR

Experimental procedures
Bacterial strains
The bacterial strains were E. coli JM109 (Promega, Madison, WI, USA), E. coli BL21(DE3)pLysS (Promega), E. coli
BMH71-18mutS (Promega), and E. coli MG1655 and
MH831 (MG1655folA::kan3). MG1655 and MH831 were
gifts from M Herrington (Department of Biology, Concordia University, Montreal, Canada) [40].

M. avium rDHFR
As described previously, the M. avium folA gene (accession
number AF006616) was cloned into the vector pET15b at
the NdeI and BamHI restriction sites (plasmid construct
p807), and expressed in E. coli strain BL21(DE3)pLysS as a

fusion protein containing an N-terminal His-tagged leader
sequence [4]. Functionality has been confirmed using a
standard DHFR assay [4].

Site-directed mutagenesis
The p807 plasmid construct of the M. avium folA gene
was used as template DNA for oligonucleotide-based sitedirected mutagenesis of Asp31 and Leu32, as well as the
control mutation Val76, using the GeneEditor Protocol
Kit (Promega). Mutagenic oligonucleotides were designed
in accordance with the Promega GeneEditor Kit recommendations (Promega TM 047), and were obtained from

3294

Integrated DNA Technologies (Coralville, IA, USA)
(Table 1). The mutagenesis reaction was accomplished
with the Promega GeneEditor Kit protocol, using the
kit’s selection oligonucleotide Top Strand. For each
mutation reaction, at least five isolated colonies were
selected and grown separately overnight in 10 mL of
LB broth with 100 lgỈmL)1 carbenicillin and 50 lL of
GeneEditor Antibiotic Selection Mix in a shaking incubator at 37 °C and 225 r.p.m. (12–14 h). Cells were harvested by centrifugation at 4 °C and maximum speed 2135 g
for 10 min (Sorval RTH-250 rotor). Plasmid DNA was
extracted using the Wizard Plus SV miniprep DNA purification system (Promega). The plasmid DNA concentration was determined in a Gene Quant proRNA ⁄ DNA
calculator (Amersham Biosciences, Piscataway, NJ, USA),

using quartz microcapillaries. Mutations were confirmed
by sequencing the full length of the folA gene insert of
the pET15b vector using T7 forward primers. Sequencing
was performed at either the Recombinant DNA ⁄ Protein
Resource Facility of Oklahoma State University (Stillwater, OK, USA) or at the Oklahoma Medical Research
Foundation (Oklahoma City, OK, USA).

High Efficiency BL21(DE3)pLysS competent cells (Promega) were transformed with p807 with and without altered
DHFR according to the manufacturer’s instructions, and
plated on LB agar (Difco, Lawrence, KS, USA) containing
100 lgỈmL)1 carbenicillin and 34 lgỈmL)1 chloramphenicol.
Plates were incubated overnight at 37 °C.

Transformants were grown overnight in 10 mL of LB
broth containing the appropriate antibiotics. Cells were
then centrifuged at maximum speed (2135 g) [Sorval
RTH-250 rotor, RT-7R centrifuge, Sorval (Thermo
Fisher Scientific, Waltham, MA, USA)] at 4 °C for
10 min, and resuspended in 10 mL of fresh LB medium
with 100 lgỈmL)1 carbenicillin and 34 lgỈmL)1 chloramphenicol. Each overnight culture was used to inoculate
500 mL of LB medium containing 100 lgỈmL)1 carbenicillin and 34 lgỈmL)1 chloramphenicol. Protein expression
was performed as described previously [4].
Wild-type M. avium DHFR and DHFR proteins containing amino acid substitutions were purified according to
Barrow et al. [5], using BugBuster protein Extraction Reagent Kit (Novagen, San Diego, CA, USA). The target protein was purified using His-Bind resin (Novagen) under
nondenaturing conditions, as described previously [4]. All

purification steps were evaluated by SDS ⁄ PAGE [41], and
protein concentrations were determined with the Bio-Rad
assay. The column was connected to the Biologic LP Chromatography System (BioRad, Hercules, CA, USA), and the
automated procedure was started. The column was washed

FEBS Journal 274 (2007) 3286–3298 ª 2007 The Authors Journal compilation ª 2007 FEBS


R. A. Bock et al.
ă

with 5 mL of buffer A (5 mm imidazole, 500 mm NaCl,

20 mm Tris ⁄ HCl, pH 7.9) at a flow rate of 1 mLỈmin)1.
Elution was continued as a linear gradient of 0–100%
buffer B (5–500 mm imidazole) over a volume of 20 mL.
Elution was continued with an additional 20 mL of 100%
buffer B (500 mm imidazole) to ensure that all protein was
eluted. Imidazole was removed by dialyzing against buffer
without imidazole and also with PD 10 columns (Sephadex
R G-25M) (Amersham Biosciences) as described previously
[5]. Upon completion, samples were removed and aliquoted
for protein determination, SDS ⁄ PAGE, and enzyme assays.
Samples for protein determination and SDS ⁄ PAGE were
stored at ) 20 °C, and samples for enzyme assay were

stored at ) 80 °C.

CD
CD spectroscopy was performed with a Jasco-715 (Jasco
Corporation, Tokyo, Japan) spectropolarimeter using a
0.1-cm path length cell over the 195–260 nm range. The
spectra were acquired every 1 nm with a 2 s averaging
time per point and a 1 nm bandpass. Data were collected
at 25 °C. Quadruplicates of the spectra were averaged,
corrected for background, and smoothed. The proteins
were dissolved in 20 mm sodium phosphate and 100 mm
NaCl (pH 7.4). The concentration of protein was determined by UV absorption spectroscopy, using an extinction coefficient, at 280 nm, of 39 500 m)1Ỉcm)1. The mean

residue ellipticity (degỈcm)2Ỉdmol)1) was calculated from
the number of residues of the recombinant DHFR (200).
The secondary structure of the proteins, including regular
and distorted a-helix, regular and distorted b-sheet, turns,
and unordered structures, was estimated with the program
selcon3 using a 29-protein dataset of basic spectra [15].

Functionality of mutant rDHFR
The functionality of the recombinant mutant forms was
determined in a standard DHFR enzyme assay [4] in comparison with the wild-type rDHFR, and by their ability or
inability to restore growth of a DHFR-deficient E. coli
strain (growth complementation), also in comparison with

the wild-type rDHFR (see below).

DHFR enzyme assay
The enzyme assay for determining functionality of the
DHFR mutant forms was performed as described previously [4,5]. Each enzyme sample was measured multiple
times. A control reaction without FAH2 was used to correct for NADPH oxidation. One unit was defined as the
amount of enzyme that reduced 1 lmole of FAH2Ỉper
minute on the basis of a molar extinction coefficient of
12 300 m)1Ỉcm)1 at 340 nm [9].

Site-directed mutagenesis of M. avium DHFR

Kinetic assay
Kinetic parameters of wild-type and mutant DHFRs were
determined with the standard assay above, with the following modifications: For FAH2, the NADPH concentration
was kept constant at 100 lm, whereas the FAH2 concentration was varied from 0.35 to 5 lm. The 1 mL reaction mixture was incubated for 1 min at 30 °C, but with NADPH
and FAH2. The reaction was initiated by addition of the
enzyme, and the activity was measured, but for 1 min at
10 s reading intervals. The amount of enzyme used was the
amount that gave a linear progress curve over the 3 min
measurement during the standard assay. For NADPH, the
concentration of FAH2 was kept constant at 100 lm,
whereas that of NADPH was varied from 0.7 to 10 lm.
The reaction components were incubated as described

above, and the reaction was also enzyme initiated. The kinetic parameters Km and Vmax for FAH2 and NADPH of
the wild-type M. avium rDHFR (p807) and the Asp31 and
Leu32 mutants were determined with the nonlinear Michaelis–Menten curve-fitting program enzfitter (Biosoft,
Great Shelford, Cambridge, UK).

Protein determination
The protein concentration was determined using the Bio-Rad
microassay in a 96-well format as described previously [5].

IC50 determination
The stock solution for TMP was prepared at
10.24 mgỈmL)1 in sterile dimethylsulfoxide (Sigma, St

Louis, MO, USA) and stored at ) 20 °C. For the drug
assay, the stock solution was diluted with sterile dimethylsulfoxide to 1.024 mgỈmL)1. This working solution was
used to prepare a series of drug concentrations ranging
from 1024 to 0.01024 lgỈmL)1.
The assay is similar to the standard enzyme assay. The
1 mL enzyme reaction mixture contained 10 lL of drug in
addition to the normal components, and was incubated at
30 °C for 3 min, after which 0.1 mm FAH2 (Sigma) was
added to initiate the reaction. Activity was measured as
previously described. A control reaction using 10 lL of
dimethylsulfoxide instead of the drug was set up to determine the effect of dimethylsulfoxide on the reaction. The
effects of NADPH oxidation were taken into account as

described for the standard assay. Percentage inhibition was
calculated by determining the quotient of the activity
obtained with drug and that obtained in the reaction with
dimethylsulfoxide only, and subtracting the quotient
(obtained from dividing the activity using drug by the activity using dimethylsulfoxide) from one, then multiplying by
100 to obtain a percentage. Values from reactions for at
least four different drug concentrations were determined for

FEBS Journal 274 (2007) 3286–3298 ª 2007 The Authors Journal compilation ª 2007 FEBS

3295

Site-directed mutagenesis of M. avium DHFR

R. A. Bock et al.
ă

each enzyme sample, two above the 50% inhibition and
two below. The concentration of the drug that inhibited the
reaction by 50% (IC50) was computed using the fourparameter curve program of the kc junior software (BioTEK, Winooski, VT, USA). The SAS statistical software
(sas system for windows V8; SAS Institute, Inc., Cary,
NC, USA) was used to evaluate statistical significance.

Electrotransformation of E. coli
Electrocompetent E. coli cells (MG1655 and MH831) were
prepared according to Bio-Rad protocols, and were electroporated with 10 ng of plasmid DNA in 0.2 cm cuvettes
with a Gene Pulser apparatus (Bio-Rad) set to 25 lF,
2.5 kV, and 200 W. Cells were recovered in 1.0 mL of SOC
medium, and were then transferred to a sterile microcentrifuge tube and incubated for 1 h at 37 °C and 225 r.p.m.
(rotary aeration). For each reaction, 100 lL was plated on
LB agar containing 30 lgỈmL)1 kanamycin, 50 lgỈmL)1
thymidine and 100 lgỈmL)1 carbenicillin. The plates were
incubated overnight at 37 °C (Isotemp incubator; Fisher
Scientific, Pittsburg, PA, USA).

Growth assay with thymidine
An isolated colony was selected and grown overnight in
12.5 mL of LB medium containing 30 lgỈmL)1 kanamycin,
50 lgỈmL)1 thymidine and 100 lgỈmL)1 carbenicillin. The
overnight culture was used ( 50 lL) to inoculate 30 mL
of LB medium (30 lgỈmL)1 kanamycin, 50 lgỈmL)1 thymidine and 100 lgỈmL)1 carbenicillin) to achieve a D600 of
0.008–0.011. This set-up culture was further diluted as follows: 25 mL was taken and combined with an additional
25 mL of fresh LB medium containing 30 lgỈmL)1 kanamycin, 50 lgỈmL)1 thymidine and 100 lgỈmL)1 carbenicillin
(1 : 1 dilution) in 125 mL sterile flasks. A D600 reading was
taken for the time point t ¼ 0. This sample was serially
diluted (10)1)10)7) with sterile H2O, and 10 lL of each
dilution was plated on LB agar with the same antibiotic

conditions as above. The cultures were grown at 37 °C and
225 r.p.m. Samples were taken after 2, 4, 6 and 8 h. For
each time interval, a D600 reading was taken, and samples
were serially diluted and plated on LB agar as described
above. The agar plates were incubated at 37 °C (Isotemp
incubator’; Fischer Scientific) overnight (about 15 h), and
plates were used to determine colony-forming units. A
growth curve was constructed by plotting D600 against
time.

Complementation of growth defect of MH831
Transformants were grown on LB agar with 30 lgỈmL)1

kanamycin, 100 lgỈmL)1 carbenicillin, and 0.1 mm isopropyl thio-b-d-galactoside, but no thymidine. The same

3296

procedure was followed as described above, but without
thymidine.

Acknowledgements
The educational support for R. A. Bocks doctoral
ă
program was provided by a Namibian Government
Scholarship and Training Program (NGSTP) sponsored through the Africa-America Institute (AAI) in

New York City. This research was funded by funds
provided by NIH ⁄ NIAID grant AI-41348 (W. W. Barrow) and the Sitlington Chair in Infectious Diseases,
Oklahoma State University (W. W. Barrow). We would
like to thank Dr Allen Edmundson and Dr Phil Bourne
for helpful suggestions regarding DHFR structural
information, as well as Esther Barrow for initial assistance with the cloning and expression of the M. avium
DHFR, the enzyme and IC50 assays, and critical
reviewing of the manuscript. Special thanks go to
Dr Phil Bourne for helpful comments regarding crystallographic models, Dr Michelle Valderas for assistance
with electrotransformation of E. coli, complementation
experiments and editing of the manuscript, and
Dr Muriel Herrington for the E. coli MH831 and

MG1655 strains. Some of these results were presented
at the 43rd annual Interscience Conference on Antimicrobial Agents and Chemotherapy (Abstract F-349,
2003, Chicago, IL, USA) and the 106th General
Meeting of the American Society for Microbiology
(Abstract U-056, 2006, Orlando, FL, USA).

References
1 Villafranca JE, Howell EE, Voet DH, Strobel MS,
Ogden RC, Abelson JN & Kraut J (1983) Directed
mutagenesis of dihydrofolate reductase. Science 222,
782–788.
2 Hartman PG (1993) Molecular aspects and mechanism

of action of dihydrofolate reductase inhibitors. J Chemother 5, 369–376.
3 McCourt M & Cody V (1991) Conformational analysis
of lipophilic antifolates: crystal structure of 2-amino-4oxo-6-adamantylpteridine and a comparison of its binding to bacterial and avian dihydrofolate reductase.
J Am Chem Soc 113, 6634–6639.
4 Zywno-vanGinkel S, Dooley TP, Suling WJ & Barrow
WW (1997) Identification and cloning of the Mycobacterium avium folA gene, required for dihydrofolate reductase activity. FEMS Microbiol Lett 156, 69–78.
5 Barrow EW, Bourne PC & Barrow WW (2004)
Functional cloning of Bacillus anthracis DHFR and
confirmation of natural resistance to trimethoprim.
Antimicrob Agents Chemother 48, 4643–4649.

FEBS Journal 274 (2007) 3286–3298 ª 2007 The Authors Journal compilation ª 2007 FEBS

R. A. Bock et al.
ă

6 Suling WJ, Seitz LE, Pathak V, Westbrook L, Barrow
EW, Zywno-van-Ginkel S, Reynolds RC, Piper JR &
Barrow WW (2000) Antimycobacterial activity of 2,4-diamino-5-deazapteridine derivatives and effects on mycobacterial dihydrofolate reductase (DHFR). Antimicrob
Agents Chemother 44, 2784–2793.
7 Kharkar PS & Kulkarni VM (2003) A proposed model
of Mycobacterium avium complex dihydrofolate
reductase and its utility for drug design. Org Biomol

Chem 1, 1315–1322.
8 Li R, Sirawaraporn R, Chitnumsub P, Sirawaraporn W,
Wooden J, Athappilly F, Turley S & Hol WGJ (2000)
Three-dimensional structure of M. tuberculosis dihydrofolate reductase reveals opportunities for the design of
novel tuberculosis drugs. J Mol Biol 295, 307–323.
9 Baccanari DP, Stone D & Kuyper L (1981) Effect of a
single amino acid substitution on Escherichia coli dihydrofolate reductase catalysis and ligand binding.
J Biol Chem 256, 1738–1747.
10 Matthews DA, Alden RA, Bolin JT, Freer ST, Hamlin
R, Xuong N, Kraut J, Poe M, Williams M & Hoogsteen K (1977) Dihydrofolate reductase: x-ray structure
of the binary complex with methotrexate. Science 197,
452–455.

11 Howell EE, Villafranca JE, Warren MS, Oatley SJ &
Kraut J (1986) Functional role of aspartic acid-27 in
dihydrofolate reductase revealed by mutagenesis.
Science 231, 1123–1128.
12 Bolin JT, Filman DJ, Matthews DA, Hamlin RC &
Kraut J (1982) Crystal structures of Escherichia coli and
Lactobacillus casei dihydrofolate reductase refined at 1.7
A resolution. I. General features and binding of methotrexate. J Biol Chem 257, 13650–13662.
13 Matthews DA, Alden RA, Bolin JT, Filman DJ,
Freer ST, Hamlin R, Hol WG, Kisliuk RL, Pastore
EJ, Plante LT et al. (1978) Dihydrofolate reductase
from Lactobacillus casei. X-ray structure of the

enzyme methotrexate.NADPH complex. J Biol Chem
253, 6946–6954.
14 Suling WJ, Reynolds RC, Barrow EW, Wilson LN,
Piper JR & Barrow WW (1998) Susceptibilities of
Mycobacterium tuberculosis and Mycobacterium avium
complex to lipophilic deazapteridine derivatives, inhibitors of dihydrofolate reductase. J Antimicrob Chemother
42, 811–815.
15 Sreerama N, Venyaminov SY & Woody RW (2000)
Estimation of protein secondary structure from circular
dichroism spectra: inclusion of denatured proteins with
native proteins in the analysis. Anal Biochem 287, 243–
251.

16 Bystroff C, Oatley SJ & Kraut J (1990) Crystal structures of Escherichia coli dihydrofolate reductase: the
NADP+ holoenzyme and the folate. NADP+ ternary
complex. Substrate binding and a model for the transition state. Biochemistry 29, 3263–3277.

Site-directed mutagenesis of M. avium DHFR

17 Ohmae E, Sasaki Y & Gekko K (2001) Effects of
five-tryptophan mutations on structure, stability and
function of Escherichia coli dihydrofolate reductase.
J Biochem (Tokyo) 130, 439–447.
18 Appleman JR, Howell EE, Kraut J & Blakley RL
(1990) Role of aspartate 27 of dihydrofolate reductase

from Escherichia coli in interconversion of active and
inactive enzyme conformers and binding of NADPH.
J Biol Chem 265, 5579–5584.
19 Birdsall B, Andrews J, Ostler G, Tendler SJ, Feeney J,
Roberts GC, Davies RW & Cheung HT (1989) NMR
studies of differences in the conformations and dynamics of ligand complexes formed with mutant dihydrofolate reductases. Biochemistry 28, 1353–1362.
20 Dunn SM, Lanigan TM & Howell EE (1990) Dihydrofolate reductase from Escherichia coli: probing the role
of aspartate-27 and phenylalanine-137 in enzyme conformation and the binding of NADPH. Biochemistry 29,
8569–8576.
21 Howell EE, Booth C, Farnum M, Kraut J & Warren
MS (1990) A second-site mutation at phenylalanine-137
that increases catalytic efficiency in the mutant aspartate-27–serine Escherichia coli dihydrofolate reductase.

Biochemistry 29, 8561–8569.
22 Thillet J, Adams JA & Benkovic SJ (1990) The kinetic
mechanism of wild-type and mutant mouse dihydrofolate reductases. Biochemistry 29, 5195–5202.
23 Polshakov VI (2001) Dihydrofolate reductase: structural
aspects of mechanisms of enzyme catalysis and inhibition. Russian Chem Bull 50, 1733–1751.
24 Adams J, Johnson K, Matthews R & Benkovic SJ
(1989) Effects of distal point-site mutations on the binding and catalysis of dihydrofolate reductase from
Escherichia coli. Biochemistry 28, 6611–6618.
25 Appleman JR, Howell EE, Kraut J, Kuhl M & Blakley
RL (1988) Role of aspartate 27 in the binding of methotrexate to dihydrofolate reductase from Escherichia
coli. J Biol Chem 263, 9187–9198.
26 Chen JT, Mayer RJ, Fierke CA & Benkovic SJ (1985)

Site-specific mutagenesis of dihydrofolate reductase
from Escherichia coli. J Cell Biochem 29, 73–82.
27 Fierke CA & Benkovic SJ (1989) Probing the functional
role of threonine-113 of Escherichia coli dihydrofolate
reductase for its effect on turnover efficiency, catalysis,
and binding. Biochemistry 28, 478–486.
28 Mayer RJ, Chen JT, Taira K, Fierke CA & Benkovic
SJ (1986) Importance of a hydrophobic residue in binding and catalysis by dihydrofolate reductase. Proc Natl
Acad Sci USA 83, 7718–7720.
29 Murphy DJ & Benkovic SJ (1989) Hydrophobic interactions via mutants of Escherichia coli dihydrofolate
reductase: separation of binding and catalysis. Biochemistry 28, 3025–3031.
30 Taira K & Benkovic SJ (1988) Evaluation of the

importance of hydrophobic interactions in drug

FEBS Journal 274 (2007) 3286–3298 ª 2007 The Authors Journal compilation ª 2007 FEBS

3297


Site-directed mutagenesis of M. avium DHFR

31

32

33

34

35

R. A. Bock et al.
ă

binding to dihydrofolate reductase. J Med Chem 31,
129–137.

David CL, Howell EE, Farnum MF, Villafranca JE,
Oatley SJ & Kraut J (1992) Structure and function of
alternative proton-relay mutants of dihydrofolate reductase. Biochemistry 31, 9813–9822.
Basran J, Casarotto MG, Barsukov IL & Roberts GC
(1995) Role of the active-site carboxylate in dihydrofolate reductase: kinetic and spectroscopic studies of the
aspartate 26–>asparagine mutant of the Lactobacillus
casei enzyme. Biochemistry 34, 2872–2882.
Cody V, Luft JR, Ciszak E, Kalman TI & Freisheim
JH (1992) Crystal structure determination at 2.3 A of
recombinant human dihydrofolate reductase ternary
complex with NADPH and methotrexate-gamma-tetrazole. Anticancer Drug Des 7, 483–491.
Falzone CJ, Wright PE & Benkovic SJ (1991) Evidence

for two interconverting protein isomers in the methotrexate complex of dihydrofolate reductase from
Escherichia coli. Biochemistry 30, 2184–2191.
Prendergast NJ, Appleman JR, Delcamp TJ, Blakley
RL & Freisheim JH (1989) Effects of conversion of
phenylalanine-31 to leucine on the function of human
dihydrofolate reductase. Biochemistry 28, 4645–4650.

3298

36 Tsay JT, Appleman JR, Beard WA, Prendergast NJ,
Delcamp TJ, Freisheim JH & Blakley RL (1990) Kinetic
investigation of the functional role of phenylalanine-31

of recombinant human dihydrofolate reductase. Biochemistry 29, 6428–6436.
37 Huang Z, Wagner CR & Benkovic SJ (1994) Nonadditivity of mutational effects at the folate binding site of
Escherichia coli dihydrofolate reductase. Biochemistry
33, 11576–11585.
38 Wagner CR, Thillet J & Benkovic SJ (1992)
Complementary perturbation of the kinetic
mechanism and catalytic effectiveness of dihydrofolate
reductase by side-chain interchange. Biochemistry 31,
7834–7840.
39 Chunduru SK, Cody V, Luft JR, Pangborn W, Appleman JR & Blakley RL (1994) Methotrexate-resistant
variants of human dihydrofolate reductase. Effects of
Phe31 substitutions. J Biol Chem 269, 9547–9555.

40 Herrington MB & Chirwa NT (1999) Growth properties
of a folA null mutant of Escherichia coli K12. Can J
Microbiol 45, 191–200.
41 Laemmli UK (1970) Cleavage of structural proteins
during the assembly of the head of bacteriophage T4.
Nature 227, 680–685.

FEBS Journal 274 (2007) 3286–3298 ª 2007 The Authors Journal compilation ª 2007 FEBS