(Received for publication, July 10, 1996, and in revised form, October 19, 1996)
From the School of Chemistry and Biochemistry, The Georgia Institute of Technology, Atlanta, Georgia 30332
The nucleotide sequence of the Shigella boydii dgt gene, which encodes the enzyme deoxyguanosine triphosphate triphosphohydrolase (dGTPase, EC 3.1.5.1), has been determined. The 1515-nucleotide Shigella dgt open reading frame has been subcloned into a T7 RNA polymerase-based expression vector. The resulting expressed protein has been purified to homogeneity using a novel single-day chromatographic regime. The protocol includes ion exchange, affinity, and hydrophobic interaction chromatography. The purified 505-amino acid (59.4 kDa) protein exists in solution as a heat-stable homotetramer, and enzymatic assays reveal that the expressed enzyme is fully active. Substrate specificity can be explained by the array of potential hydrogen bond donors/acceptors displayed on the base moiety of the (deoxy)nucleoside triphosphate. Shigella dGTPase can be inhibited by the addition of stoichiometric amounts of reducing agents. The loss of activity is both time- and concentration-dependent and is accompanied by a decrease in the thermal stability of the enzyme. Shigella dGTPase in the fully reduced form is destabilized by 1.8 kcal/mol compared with the oxidized form. Hence, disulfide bonds play a pivotal role in the maintenance of dGTPase stability and enzymatic functionality. Initial Shigella dGTPase protein crystals have been formed.
Deoxyguanosine triphosphate triphosphohydrolase (dGTPase1; EC 3.1.5.1) was discovered by Kornberg et al. (1) during the purification of DNA polymerase I. The enzyme was partially purified and found to catalyze the hydrolysis of deoxyguanosine triphosphate to deoxyguanosine and inorganic tripolyphosphate:
![]() |
![]() |
Seto et al. (2) demonstrated that dGTPase was overproduced by approximately 20-30-fold in a mutant strain of E. coli previously described by Saito and Richardson (3). This observation was also made by Beauchamp and Richardson (4) who purified the enzyme from these bacteria and showed that it was identical to the dGTPase originally characterized. Huber et al. (5) and Nakai and Richardson (6) demonstrated that bacteriophage T7 produce a specific dGTPase inhibitor protein. The structural gene encoding E. coli dGTPase (dgt) was first isolated by Quirk et al. (7), and the sequence of the E. coli dgt gene was reported by Wurgler and Richardson (8) and Quirk et al. (9).
The dGTPase enzyme is specific to the bacterial family Enterobacteriaceae and is expressed in all genera of the family with the exception of Erwinia species (10). Wurgler and Richardson (11) were the first to demonstrate that dGTPase binds ssDNA in a cooperative manner at physiological salt concentrations. The significance of these observations is unknown.
We are interested in understanding the enzyme family from an evolutionary perspective and have embarked on a project to clone and study all of the enteric dgt homologues. Comparative enzymological work has been frustrated in part by the difficulty in producing homogeneous dGTPase from native or recombinant sources. To aid in our continuing work on the Enterobacteriaceae dGTPases, we report a rapid, streamlined, and automated purification protocol. This purification scheme is employed to isolate homogeneous and active recombinant dGTPase from a second member of the Enterobacteriaceae, Shigella boydii. This material has been utilized in a variety of biochemical and biophysical experiments to further the understanding of this novel family of enzymes.
Preparation of plasmid DNA, restriction enzyme digests, ligation, and agarose gel electrophoresis were performed according to Maniatis et al. (12) or Sambrook et al. (13). Bacterial transformation was as described by Morrison (14). Protein SDS-PAGE gels were made, run, and processed as per Laemmli (15). Protein concentration was determined according to the method of Bradford (16) using bovine serum albumin as a standard. Chemical reagents were from Sigma, except for deoxynucleoside triphosphates which were purchased from Pharmacia Biotech Inc. S. boydii was purchased from the American Type Culture Collection (ATCC 9207). Vent DNA polymerase used in all PCR reactions was purchased from New England Biolabs and used with the supplied buffer.
dGTPase AssaysThe procedure measures the hydrolysis of
[-32P]dGTP to Norit nonadsorbable 32P and
is performed as described by Seto et al. (2). In order to
measure the rate of hydrolysis of nonradioactive substrates, a
colorimetric assay was performed in conjunction with an assay of
orthophosphate (Pi) according to Ref. 17 after hydrolyzing reaction products in 0.5 N HCl for 15 min in a boiling
water bath.
Genomic DNA was isolated from an overnight culture of
S. boydii using the Genomic DNA Prep kit from Bio101, Inc.
The Shigella dgt gene was amplified from this genomic DNA
template using degenerate PCR oligonucleotide primers based upon the
amino acid sequence of Escherichia dGTPase. Oligonucleotide
A had the sequence 5-ATGGCNCA(A/G)ATNGA(T/C)TT(T/C)(C/A)GNA and
oligonucleotide B had the sequence 5
-A(T/C)TG(T/C)NACNGCCATNA(A/G)NC, where N is an equimolar mixture of all four bases. The PCR product was
concentrated by ethanol precipitation and cloned into
SmaI-digested, dephosphorylated pUC18. The ligated material
was used to transform competent HB101 bacteria. Recombinant colonies
were screened for the presence of full-length inserts by miniprep
analysis and restriction digestion coupled with agarose gel
electrophoresis analysis of digestion products. The plasmid constructs
were designated pSdgt1,2. Double-stranded DNA sequencing was performed
by the dideoxy termination method (18) using Sequenase version 2.0 (U. S. Biochemical Corp.). Additional DNA sequencing was performed using
the Applied Biosystems, Inc. model 373A automatic DNA Sequencer. The
separate isolates were fully sequenced in order to check for
PCR-induced errors.
The expression plasmid pSdgtE was created using PCR-directed
mutagenesis of pSdgt. A unique NcoI site was introduced at
the initiation ATG codon using the oligonucleotide
5-CTCGCCCAATGGCACAGATTGATTTCCGA-3
. Similarly, a unique
BamHI site was introduced at the 3
end of the coding
sequence immediately downstream from the TAA termination codon using
the oligonucleotide 5
-TCGTACGGATCCTTATTGTTCTACGGCCATCA-3
. The final
PCR product was ethanol-precipitated, digested with NcoI and
BamHI, and ligated into the expression vector pET11d (Novagen, Inc.) which had been digested with NcoI,
BamHI and dephosphorylated. The sequence of the expression
construct was verified by DNA sequencing.
Plasmid pSdgtE containing BL21(DE3)-pLysS cells were
grown at 37 °C in Luria broth supplemented with 60 µg/ml ampicilin
from a 1% inoculum.
Isopropyl-1-thio--D-galactopyranoside was added to a
final concentration of 1 mM when the cells had reached an A595 value of 0.8 (in approximately 3 h
postinoculation). Cell growth continued for 5 additional hours before
harvesting. Typically, 3 g of cells were obtained per liter.
Cells were pelleted by centrifugation at 10,000 × g for 10 min and resuspended in 1 volume of 10 mM Tris-HCl, pH 8.0. The cells were respun as above and resuspended in 2 volumes of 10 mM Tris, pH 8.0, 1 mM EDTA, 1 mM sodium azide (Buffer I). The cells were disrupted by sonication on ice using a Branson sonicator equipped with a microtip. The extract was clarified by centrifugation at 12,000 × g for 20 min, and the supernatant was designated as Fraction I. All subsequent chromatography steps were performed using a Bio-Rad Econo System in automatic peak detection/sample loading mode. The purification protocol is designed so that the pooled fraction from one step can be automatically loaded onto the next column. During the pilot runs, enzyme assays were performed on each fraction in order to determine where active dGTPase was eluting, followed by SDS-PAGE analysis. Column chromatograms were so reproducible that automated peak pooling could be employed during production runs. This effectively saved the time required to perform enzyme assays on individual fractions and visualization of individual fractions by SDS-PAGE.
Fraction I was applied to a 4.9 cm2 × 20 cm column of DEAE-Sepharose (Sigma) in Buffer I. The enzyme was eluted from the column with a linear gradient of 0 to 0.5 M KCl at a flow rate of 2.0 ml/min. The major peak containing dGTPase activity eluted as a broad peak centered at 0.3 M KCl. The central 90% of the absorbance was automatically pooled and constituted Fraction II.
Fraction II was immediately applied to a 1.8 cm2 × 10 cm single-stranded DNA-cellulose column (Sigma) in Buffer I, 300 mM KCl. The column was washed with 120 ml of the Buffer I, followed by a linear 50-ml KCl gradient from 2.0 M to 4.0 M KCl. The flow rate throughout this elution was 2.0 ml/min. The enzyme eluted from the column halfway through the gradient. The central 95% of the eluting material was automatically pooled and was designated as Fraction III.
Homogeneous Shigella dGTPase was prepared from Fraction III via hydrophobic interaction chromatography. Ammonium sulfate (from a 4 M stock solution) was added to the Fraction III pool to a final concentration of 1.0 M. This material was applied to a Bio-Rad Econo-Pac t-butyl hydrophobic interaction chromatography cartridge column (1 ml total column volume). A 100-ml reverse concave gradient from 1.0 M (NH4)2SO4, 3.0 M KCl, Buffer I to Buffer I alone was applied directly after loading the sample. Homogeneous dGTPase eluted from the column during the last third of the gradient. The central 95% of the peak (as measured by absorbance) was automatically pooled. The material was concentrated to a final volume of 2 ml by pressure filtration through a semipermeable membrane (Amicon YM-3) and dialyzed versus 2 liters of 10 mM Tris-HCl (pH 7.5). The concentrated/dialyzed material constituted Fraction IV. All studies were performed with homogeneous (Fraction IV) dGTPase.
Intrinsic Tryptophan FluorescenceStability measurements of reduced and oxidized dGTPase were performed by measuring protein unfolding in the presence of guanidine hydrochloride (GdnHCl) via intrinsic tryptophan fluorescence in a SPEX Fluorolog fluorimeter. The excitation and emission wavelengths were 295 nm and 340 nm, respectively. Both excitation and emission monochrometer slits were set at 5 nm. Protein (10 µM) was reduced with dithiothreitol (DTT) for 2 h, GdnHCl was then added to individual samples in the final concentration range from 0 to 1.5 M, and the samples were incubated at room temperature for 10 h to ensure that unfolding equilibrium had been achieved. Relative fluorescence was converted into free energy values as described previously (19).
The nucleotide and deduced amino acid sequences of Shigella dGTPase indicate a 1515-base pair open reading frame that corresponds to a protein of 505 amino acid residues, with a calculated molecular mass of 59.4 kDa. The Shigella dgt gene sequence differs from the Escherichia homologue in 13 positions. Seven of the base changes occur in the third codon position, and the remaining six nucleotide changes are evenly distributed between first and second codon positions. These nucleotide substitutions give rise to only six amino acid differences between the Escherichia and Shigella enzymes (98.8% identity). Most of the amino acid differences are conservative substitutions. Codon usage is typical of other Shigella genes. The nucleotide sequence has been reported to GenBankTM and assigned accession number U42434[GenBank].
Overproduction and Isolation of Recombinant Shigella dGTPaseThe Shigella dgt open reading frame
was subcloned into the protein expression vector pET11d (Novagen, Inc.)
using PCR primers that incorporated a 5-NcoI site and a
3
-BamHI site in order to facilitate in frame directional
cloning. Bacteria transformed with the expression construct pSdgtE
synthesize high levels of soluble dGTPase. The specific activity of
crude protein extracts prepared from BL21(DE3)pLysS bacteria harboring
pSdgtE indicates that the material is overproduced 210-fold compared
with the native expression level as shown in Table I.
DEAE-Sepharose chromatography increases the specific activity of the
protein preparation 5.6-fold, but primarily serves to remove the ssDNA
that remains associated with the enzyme in crude
extract.2 This material is then applied to
a ssDNA-cellulose column in an affinity purification step that results
in nearly homogeneous dGTPase. The final purification step on a
t-butyl hydrophobic interaction chromatography column
efficiently removes the last few remaining contaminant bands. The
resulting homogeneous enzyme preparation represents a 232-fold
purification from the crude bacterial extract with a 52% recovery.
Overproduction of dGTPase, coupled with this purification protocol
yields 6.4 mg of protein per 25 g of cell paste. Edman degradation
of the N-terminal region of the enzyme (data not shown) confirms the
identity of the overexpressed material as bona fide dGTPase and reveals
that the initiating methionine residue is removed by E. coli
as a post translational modification. The entire purification scheme
(from the preparation of the crude extract to homogeneous enzyme) takes
approximately 8 h to complete.
|
The
reaction conditions for optimal dGTPase activity are similar to the
E. coli protein (2, 4, 10). The enzyme has a broad pH
optimum profile that is centered at a pH value of 8.2, and activity is
absolutely dependent on the presence of the divalent cation magnesium.
Like its E. coli counterpart, the Shigella enzyme is not active when other divalent cations are substituted for Mg2+. Shigella dGTPase has Km
of 10 µM (Vmax 1.5 µmol
min1 mg
1) for dGTP and a
Km of 120 µM for GTP
(Vmax 0.06 µmol min
1
mg
1). Analytical gel filtration experiments (20)
indicate that purified dGTPase is a tetramer with a 52.8-Å Stokes
radius and a frictional coefficient of 1.12 (all data not shown).
Substrate specificity was determined for the enzyme by measuring the
ability of dGTPase to hydrolyze several other deoxynucleoside triphosphates. Fig. 1 indicates that of the four
canonical dNTPs, only dGTP is hydrolyzed to a significant extent (and
arbitrarily set to 100% activity), dTTP shows a 20% hydrolysis rate
relative to dGTP, and neither dATP nor dCTP is hydrolyzed. In order to expand the understanding of how dGTPase discriminates between substrate
and nonsubstrate, two additional molecules were assayed. Both dITP and
dUTP are hydrolyzed by dGTPase at 40% and 25% of the dGTP rate,
respectively.
Shigella dGTPase is sensitive to the action of reducing
agents. Dithiothreitol preincubated with 0.4 µM dGTPase
results in a concentration-dependent loss of enzymatic
activity (Fig. 2). This loss of activity occurs at molar
ratios ranging from 100:1 to 2500:1 (DTT to dGTPase). The addition of
substrate to the preincubation mixture does not prevent the loss of
enzymatic activity (data not shown). In addition, a time course
experiment was undertaken that shows that the reduction rate at a DTT
concentration of 300 µM is time-dependent, as
one would expect for the reduction of disulfide bonds (Fig.
3). This experiment also indicates that the loss of
enzymatic activity is due to the phenomenon of disulfide reduction and
not a general effect of DTT interfering with the assay. Fig. 3 shows
that enzymatic activity decays with respect to time following
pseudo-first order kinetics with a rate constant of 3.3 × 104 s
1 producing a half-life of 35 min.
There is a residual 12% of total activity remaining after 120 min of
preincubation in the presence of the reductant. Reduction of dGTPase
with 2-mercaptoethanol resulted in similar loss of activity kinetics
(data not shown). An oxidized enzyme control (Fig. 3) indicates that
native dGTPase is active over the time course of the assay.
Oxidized Shigella dGTPase remains fully enzymatically active
at 60 °C for extended incubation periods (Fig. 4).
This thermostability is not evidenced in reduced dGTPase. To measure
the thermostability of reduced dGTPase, the enzyme (0.4 µM) is incubated with 300 µM DTT to a 60%
activity level relative to the oxidized protein (incubation time is 10 min). The reduced protein is then diluted with 10 mM
Tris-HCl (pH 7.5) to a final concentration of 0.04 µM
prior to a time course incubation at 60 °C and subsequent enzyme assay. Fig. 4 indicates that the reduced protein becomes thermolabile. Reduced dGTPase shows a loss of enzymatic activity with
respect to time following pseudo-first order kinetics with a rate
constant of 6.3 × 104 s
1, producing a
half-life of 18 min. There is a residual 5% activity remaining after
40 min of incubation at 60 °C. Reduction of dGTPase results in the
irreversible loss of enzymatic activity (Fig. 4), most probably due to
improper reformation of disulfide bonds (disulfide mixing).
The unfolding course of oxidized and fully reduced dGTPase at a
concentration of 12 µM was followed by intrinsic
tryptophan fluorescence as shown in Fig. 5. Analysis of
the unfolding curves indicates that oxidized dGTPase has a native free
energy, GH2O, of 3.7 kcal/mol, and shows a midpoint in the unfolding reaction ([GdnHCl]1/2) of 0.95 M GdnHCl. Reduced dGTPase
is significantly less stable than the oxidized protein (Fig. 5) as
evidenced by a
GH2O value
of 1.9 kcal/mol, and a [GdnHCl]1/2 value of 0.65 M GdnHCl. This results in a free energy difference
(
G) between oxidized and reduced dGTPase of
1.8 kcal/mol.
Crystallization of Shigella dGTPase
Initial crystallization
conditions were screened by the vapor diffusion method (21) using the
sparse matrix sampling techniques (22) with 5 µl of the concentrated
and dialyzed protein equilibrated against an equal volume of the
reservoir solution. Rectangular, plate-like crystals formed in 5-10
days at room temperature from a reservoir solution containing
2-methyl-2,4-pentanediol (MPD) and sodium chloride in sodium acetate
buffer. In order to optimize crystal growth, wide ranging titrations of
MPD, sodium chloride, and sodium acetate concentration were undertaken.
Crystals were reproducibly obtained from a reservoir containing 30%
MPD, 150 mM sodium chloride, 200 mM sodium
acetate (pH 5.0). Protein crystals obtained typical dimensions of
1.0 × 0.1 × 0.02 mm as shown in Fig. 6.
These crystals diffract to only 4.5 Å resolution (data not shown).
Trials to improve crystal quality are in progress.
The Shigella and Escherichia dGTPase enzymes
are highly conserved. This is to be expected, since the two genera are
the most closely related of all the Enterobacteriaceae. An analysis of the dgt gene from a second species within the genus
Escherichia, namely E. fergusonii (ATCC 35469),
reveals that there are only two nucleotide substitutions
(T381 G and G918
C; both at the third
codon position). This gives rise to only one amino acid substitution
(Phe137
Leu), indicating a high level of intergenus
sequence conservation (data not shown, but submitted to
GenBankTM with the accession number U42435[GenBank]). Both
homologues contain the previously described dGTPase amino acid
consensus structures (8, 9, 10). So, generally speaking, the accepted
taxonomy of enteric bacteria (23) is reflected in the divergence of
dGTPase sequences. We are in the process of cloning dgt
homologues from more distantly related Enterobacteriaceae genera. We
hope to learn more about the nature of the dGTPase proteins by an
extensive analysis of sequence divergence.
The expression construct pSdgtE, produces a significant amount of Shigella dGTPase that is readily identified in crude bacterial extracts. The enzyme purification scheme outlined in this paper produces homogeneous material in 1 day. This represents a significant improvement over traditional purification methods (2, 4). This overproduction level, as measured by the specific activity of induced crude protein extract, is directly comparable to the construct that overproduces E. coli dGTPase that was described by Wurgler and Richardson (11). Both specific activity measurements probably represent the maximum amount of dGTPase that can be overproduced in a T7 polymerase-based expression system without deleterious effects to the host bacterium (meaning the overproduced dGTPase completely hydrolyzes the cellular dGTP pool). The choice of bacterial host strain was also important to the success of the expression scheme. BL21(DE3) cells produced lower amounts of dGTPase than did BL21(DE3)-pLysS when transformed with pSdgtE. This is presumably due to the tighter repression of T7 RNA polymerase afforded by pLysS strains prior to induction. The key to the success of this purification protocol lies with the chromatographic reproducibility afforded by the Bio-Rad Econo System. This reproducibility permits automatic peak pooling, which in turn cuts down on the time needed to assay individual fractions for dGTPase. The chromatographic regime is also designed to permit direct pool loading onto the next column, thus avoiding dialysis buffer changes. The overexpression level, coupled with the purification protocol, produces more purified dGTPase in 1 day, than was produced in 1.5 weeks using other purification schemes (2, 4, 8, 9, 10).
The substrate specificity of dGTPase can be directly correlated to a pattern of Watson-Crick type hydrogen bonding surfaces that display the arrangement of hydrogen bond acceptor (6-position carbonyl of the purine ring), hydrogen bond donor (1-position H of the purine ring), and hydrogen bond donor (2-position amino group of the purine ring). This work indicates that the most important group of the three for determining activity is the 6-position carbonyl group. Substitution of that group with a potential hydrogen bond donor (a 6-position exocyclic amino group), as in dATP, abolishes activity. No enzymatic activity is measured on dCTP for the same reason. In addition, the substitution of the 2-position exocyclic amino group (as in dGTP) with a carbonyl (as in dCTP) changes the polarity of potential H bond interactions. Hence, dCTP has no potential interactions that are analogous to those in dGTP. The pyrimidines dUTP and dTTP show similar activities when assayed as substrates. Activity is not totally abolished, however, because the 6-position carbonyl and the 1- position H are still in place helping to lock the molecule in the active site. This also indicates that the nature of the substituent at the 2-position is not a paramount determinant for dGTPase specificity. Although simple elimination of the 2-position exocyclic amino group as in dITP results in a 60% loss of activity. Substrate discrimination based on the array of hydrogen bonding groups of the purine base is unique and is not duplicated in any studied nucleotide metabolizing enzyme (for a review see Ref. 24). This unusual specificity may be a clue to the function of the enzyme, provides the first insight into the dGTPase active site, and provides a logical rationale for the choice of other substrates/substrate analogs for further study.
In order to determine whether or not any of the 24 cysteine residues found in the dGTPase tetramer are involved in disulfide bridges, and whether or not such disulfide bonds are important to the enzymatic activity and stability of the protein, a series of reduction experiments was undertaken. The exact number and location of disulfide bonds in Shigella dGTPase is being investigated, yet clearly disulfide bonds are an integral component to the stability and biological activity of dGTPase.
It has been shown that dGTPase is not essential for bacterial viability (7, 8, 9) under laboratory conditions, but the physiological role of the enzyme may only be manifested in less controlled environments. It may also be possible that dGTPase is required during anaerobic growth, during enteropathogenesis, or in conjunction with secondary nucleotide biosynthetic mutations. We are endeavoring to elucidate the physiological role of this enzyme in addition to biophysical and evolutionary characterization. So it is hoped that by approaching the study of dGTPase study from many scientific disciplines the role of this enigmatic protein will ultimately yield to scrutiny. Certainly a streamlined purification protocol, an explanation of substrate specificity rules, and an analysis of protein behavior in reducing environments is a step in that direction.
We thank Dr. M. J. Bessman of the Johns Hopkins University for the gift of bacterial strains and purification advice. Anne Handley and Scott Chasse are thanked for critical reading of the manuscript, and Dr. John Haseltine is appreciated for his thoughtful discussions of phosphoester hydrolysis.