Class II ribonucleotide reductases catalyze carbon-cobalt bond reformation on every turnover

Article abstract of DOI:10.1021/ja9913840

Ribonucleoside triphosphate reductase (RTPR) from Lactobacillus leichmannii catalyzes the reduction of nucleotides to deoxynucleotides with concomitant oxidation of two cysteines within the active site to a disulfide. RTPR requires adenosylcobalamin (AdoCbl) as a cofactor and as a radical chain initiator. Catalysis is initiated by homolysis of the carbon-cobalt bond of AdoCbl to yield cob(II)alamin, 5'-deoxyadenosine, and a protein-based thiyl radical. The turnover numbers for ATP with its allosteric effector dGTP, and for CTP with its allosteric effector dATP, are both 2 s^-1. The rate- limiting step for turnover in the steady state is re-reduction of the oxidized form of the protein, a conformational change, or both. Under conditions where [RTPR] >> [AdoCbl], the rates of ATP and CTP reduction do not vary linearly with [AdoCbl] but instead exhibit saturation behavior with turnover numbers of 10 s^-1 (ATP) and 8.5 s^-1 (CTP). This result suggests that dissociation of AdoCbl, which requires carbon-cobalt bond reformation, follows nucleotide reduction, but precedes the rate-limiting step in catalysis. A presteady-state analysis of the ATP reduction (using rapid chemical quench methods) in the presence of [5'-³H]-AdoCbl reveals formation of product dATP at a k(obs) of 55 ± 10 s^-1 and tritium washout from [5'- ³H]-AdoCbl at 0.6 s^-1. The rate of washout is approximately equivalent to the rate of washout of ³H in the absence of substrate. Measurement of the ratio of ³H₂O:dATP over time reveals that washout of ³H occurs at the end of each turnover. Production of ³H₂O requires reformation of the carbon- cobalt bond. These steady-state and presteady-state data suggest that carbon- cobalt bond reformation and dissociation of AdoCbl into solution accompany each turnover and that the radical chain length of the RTPR-catalyzed nucleotide reduction is approximately one.

Full text of DOI:10.1021/ja9913840

VOLUME 121, NUMBER 33
AUGUST 25, 1999
©
Copyright 1999 by the
American Chemical Society
Class II Ribonucleotide Reductases Catalyze Carbon-Cobalt Bond
Reformation on Every Turnover
†
Stuart S. Licht, Christopher C. Lawrence, and JoAnne Stubbe*
Contribution from the Departments of Chemistry and Biology, Massachusetts Institute of Technology,
Cambridge, Massachusetts 02139
ReceiVed April 27, 1999
Abstract: Ribonucleoside triphosphate reductase (RTPR) from Lactobacillus leichmannii catalyzes the reduction
of nucleotides to deoxynucleotides with concomitant oxidation of two cysteines within the active site to a
disulfide. RTPR requires adenosylcobalamin (AdoCbl) as a cofactor and as a radical chain initiator. Catalysis
is initiated by homolysis of the carbon-cobalt bond of AdoCbl to yield cob(II)alamin, 5′-deoxyadenosine,
and a protein-based thiyl radical. The turnover numbers for ATP with its allosteric effector dGTP, and for
-
1
CTP with its allosteric effector dATP, are both 2 s . The rate-limiting step for turnover in the steady state is
re-reduction of the oxidized form of the protein, a conformational change, or both. Under conditions where
[
RTPR] . [AdoCbl], the rates of ATP and CTP reduction do not vary linearly with [AdoCbl] but instead
-
1
-1
exhibit saturation behavior with turnover numbers of 10 s (ATP) and 8.5 s (CTP). This result suggests
that dissociation of AdoCbl, which requires carbon-cobalt bond reformation, follows nucleotide reduction,
but precedes the rate-limiting step in catalysis. A presteady-state analysis of the ATP reduction (using rapid
chemical quench methods) in the presence of [5′- H]-AdoCbl reveals formation of product dATP at a kobs of
5
3
-1
3
-1
5 ( 10 s and tritium washout from [5′- H]-AdoCbl at 0.6 s . The rate of washout is approximately equivalent
3
3
to the rate of washout of H in the absence of substrate. Measurement of the ratio of H2O:dATP over time
3
3
reveals that washout of H occurs at the end of each turnover. Production of H2O requires reformation of the
carbon-cobalt bond. These steady-state and presteady-state data suggest that carbon-cobalt bond reformation
and dissociation of AdoCbl into solution accompany each turnover and that the radical chain length of the
RTPR-catalyzed nucleotide reduction is approximately one.
Introduction
biosynthesis, the ribonucleotide reductases (RNRs). These
enzymes utilize tyrosyl radicals, glycyl radicals, or adenosyl-
cobalamin (AdoCbl), depending on the source from which they
are isolated, to generate transient thiyl radicals that are essential
There are an increasing number of enzymes involved in
primary steps in metabolism in which metallo-cofactors act as
radical chain initiators. One important mechanistic question is
whether one initiation event results in multiple conversions of
substrates to products or whether an initiation event must occur
on every turnover. In chemical terms, one would like to
determine the chain length of the reaction. We have long been
interested in the enzymes that catalyze an essential step in DNA
1
1
in catalysis. Studies on the class II, AdoCbl-dependent,
ribonucleoside triphosphate reductases (RTPRs) are now re-
ported that provide the first direct evidence that these enzymes
catalyze, on each turnover of substrate to product, carbon-cobalt
bond homolysis and reformation. The chain-length of the
reaction is one.
†
Current address: University of California, Berkeley.
(1) Stubbe, J.; van der Donk, W. A. Chem. ReV. 1998, 98, 705-762.
Class II RNRs catalyze the conversion of nucleoside triph-
1
0.1021/ja9913840 CCC: $18.00 © 1999 American Chemical Society
Published on Web 08/06/1999

7464 J. Am. Chem. Soc., Vol. 121, No. 33, 1999
Licht et al.
Scheme 1
osphates (NTPs) to deoxynucleoside triphosphates (dNTPs). In
addition, these enzymes also catalyze the exchange of hydrogens
from the 5′ position of AdoCbl with solvent.2,3 This exchange
reaction occurs in the absence of substrate, and its rate is
maximized when a dNTP allosteric effector is present and the
Scheme 1 provides an overview of our current working model
for these reactions.
The function of AdoCbl is to generate a thiyl radical, 5′-
deoxyadenosine (5′-dA) and cob(II)alamin: the initiation event
5
in both reactions. Studies of Tamao and Blakley and Licht and
4
enzyme is in the reduced state. Seminal experiments of Tamao
8
Stubbe have shown that the kinetics of cob(II)alamin formation
and Blakley, Orme-Johnson et al., and more recent studies from
our laboratories using stopped flow (SF) UV-vis spectroscopy,
rapid chemical quench, and rapid freeze quench (RFQ) EPR
spectroscopy, have provided much insight into the mechanism
and disappearance in the presence of a NTP substrate are
biphasic: there is an initial rapid phase in which as much as
-
1
0
.4 equiv of cob(II)alamin are formed with a kobs of >200 s ,
followed by a slower phase in which cob(II)alamin is consumed
4
-8
of the exchange reaction as well as the reduction reaction.
to a steady state level of 0.1 equiv with a rate constant of 20-
-
1
-1
(
2) Beck, W. S.; Abeles, R. H.; Robinson, W. G. Biochem. Biophys.
Res. Commun. 1966, 25, 421-425.
3) Hogenkamp, H. P. C.; Ghambeer, R. K.; Brownson, C.; Blakley, R.
L.; Vitols, E. J. Biol. Chem. 1968, 243, 799-808.
50 s . The steady-state rate of dNTP formation is 2 s , and
thus carbon-cobalt bond cleavage and reformation are much
(
8
faster than steady-state turnover. At present, our studies and
(
4) Licht, S. S.; Booker, S.; Stubbe, J. Biochemistry 1999, 38, 1225-
earlier studies of Tamao and Blakley suggest that the rate-
limiting step is re-reduction of the disulfide bond generated
during dNTP formation and or a conformational change
associated with this process.
1
233.
(
(
5) Tamao, Y.; Blakley, R. L. Biochemistry 1973, 12, 24-34.
6) Orme-Johnson, W. H.; Beinert, H.; Blakley, R. L. J. Biol. Chem.
1
974, 249, 2338-2343.
(
7) Licht, S.; Gerfen, G. J.; Stubbe, J. Science 1996, 271, 477-481.
Our recent studies on the exchange reaction in the absence
of substrate suggest that this reaction is concerted (Scheme 1)4
(8) Licht, S. S. Ph.D. Thesis, Chapter 5. Massachusetts Institute of
Technology, 1998.

C-Co Bond Reformation Catalyzed by RTPR
J. Am. Chem. Soc., Vol. 121, No. 33, 1999 7465
or contains a high energy 5′-deoxyadenosyl radical intermediate,
and that 5′-hydrogens from the cofactor appear in the solvent
-
1 7
with a rate constant of >10 s . However, when substrate is
present, the thiyl radical is thought to abstract a hydrogen atom
from the 3′-position of the nucleotide to initiate the reduction
concomitant with formation of an active site disulfide. The thiyl
radical is then regenerated when the product 3′-deoxynucleotide
radical is reduced. At this stage, the thiyl radical could initiate
multiple rounds of nucleotide reduction without reformation of
the carbon-cobalt bond of AdoCbl. Alternatively, AdoCbl could
be reformed at the end of each turnover. Exchange of tritium
3
into solvent when [5′- H]-AdoCbl is the cofactor requires
reformation of the carbon-cobalt bond and hence can be used
as an indicator of this process.
Two types of experiments are presented that indicate that the
chain length of the reduction reaction is approximately one. In
one set of experiments, the concentration of RTPR is in large
excess over AdoCbl. The kcat for deoxynucleotide (dNTP)
formation is measured by monitoring NADPH oxidation in a
coupled assay using thioredoxin (TR) and thioredoxin reductase
Figure 1. Dependence on [AdoCbl] of the rate of the turnover of (A)
ATP (O, solid line) and (B) CTP (×, dashed line) under conditions in
which [RTPR] . [AdoCbl]. Each assay contained [RTPR] ) 10 ×
[
AdoCbl], 1 mM ATP (A) or 1 mM CTP (B), 1 mM dGTP (A) or 1
mM dATP (B), 85 µM TR, 1.5 µM TRR, 0.25 mM NADPH, 200 mM
sodium dimethylglutarate (NaDMG) pH 7.3. Product formation was
determined by monitoring NADPH consumption at 340 nm.
(TRR) (Scheme 1). These studies show that the rate of
re-reduction of the disulfide (measured by a TR/TRR/NADPH
coupled assay) generated during NTP reduction exceeds the
product of kcat and the concentration of the limiting reagent,
AdoCbl. These observations require that a single AdoCbl can
catalyze formation of multiple dNTPs and hence multiple
disulfides. AdoCbl therefore must dissociate from the oxidized
RTPR, a process which requires carbon-cobalt bond reforma-
tion. The rate of this process provides an upper limit for the
chain length of this reaction.
a coupled assay with TR, TRR, NADPH, the Km,AdoCbl and kcat
were determined for reduction of ATP and CTP. With both
-1
nucleotides, Km,AdoCbl is 0.2 µM, and kcat is 2 s . These numbers
9
are similar to those previously observed.
Kinetic Parameters Under Conditions in which [RTPR]
.
[AdoCbl] . Km,AdoCbl. Experiments were also carried out
under conditions in which the ratio of [RTPR] to [AdoCbl] was
0:1, (where [RTPR] varied from 1 to 30 µM). Under these
1
In a second set of experiments carried out in the presteady
conditions, given the Km,AdoCbl determined above, all AdoCbl
is bound to RTPR. The rate of dNTP formation was measured
spectrophotometrically using TR/TRR/NADPH as described
above. Above 0.8 µM AdoCbl, the rates were measured by SF
UV-vis spectroscopy. The results with ATP as substrate and
dGTP as allosteric effector are shown in Figure 1A. Interest-
ingly, as the [AdoCbl] is increased, the rate of dNTP formation
becomes independent of [AdoCbl]. The limiting value of
state in the presence of substrate, the exchange of tritium from
3
[
5′- H]-AdoCbl with solvent is used to monitor reformation of
the carbon-cobalt bond. Comparison of the rate of tritium
washout from the cofactor to the rate of dNTP formation
provides an additional measure of the frequency of carbon-
cobalt bond reformation. The ramifications of Nature’s choice
of a chain length of approximately one will be discussed.
-
1
ν/[AdoCbl) is 10 s and is 5-fold greater than the turnover
Results
-1
number of 2 s observed under standard assay conditions.
A similar experiment was carried out in the presence of CTP
with dATP as an allosteric effector. The results are shown in
Figure 1B. Once again, the rate constant for dCTP formation is
Steady State Analysis Indicating Carbon-Cobalt Bond
Reformation on Each Turnover. The chain length of radical
dependent reactions in biological systems has yet to be
unambiguously established in any reaction. We have developed
a method that could be generally useful in measuring chain
lengths for AdoCbl-requiring enzymes. The method is based
on the ability to use the steady-state rate of substrate turnover
to measure AdoCbl dissociation from the enzyme at the end of
each turnover. This process requires carbon-cobalt bond
reformation. We have thus measured the rate of dNTP formation
under two sets of conditions: Michaelis-Menten conditions
where [AdoCbl] . [RTPR] and conditions in which [RTPR]
-1
8.5 s , 4.5-fold higher than the turnover number observed under
standard conditions. Given the experimental design, the only
way one can observe turnover numbers greater than those
observed when enzyme is saturated with substrate is if each
AdoCbl is capable of participating in several turnovers of
nucleotide to deoxynucleotide prior to the rate-determining step
5
in nucleotide reduction. Previous studies of Tamao and Blakley
8
and more recent studies from our laboratory suggest that either
re-reduction of the active-site disulfide to two cysteines, a
conformational change, or both are rate-limiting in the steady
state. The simplest interpretation of these data is that the
.
[AdoCbl] . Km,AdoCbl. In the latter case, the [RTPR] that
will turn over substrate (holoenzyme) is equal to the concentra-
tion of AdoCbl, the limiting reagent, if AdoCbl does not
dissociate from RTPR. The rate of turnover (ν) would thus vary
directly with [AdoCbl], and thus ν/[AdoCbl] would be constant.
On the other hand, if AdoCbl dissociates more rapidly than the
rate of dNTP production in the steady state, a single AdoCbl
could effect conversion of multiple NTPs to dNTPs. It is
possible that saturation kinetics would be observed, limited by
the rate at which AdoCbl dissociates from RTPR.
-1
observed rate constant of RTPR re-reduction of ∼10 s reflects
the rate of AdoCbl dissociation from RTPR, which allows
accumulation of a pool of oxidized RTPR that is greater than
[
AdoCbl].
3
3
H2O Release from [5′- H]-AdoCbl in the Presteady State
as an Indicator of C-Co Bond Reformation. Recently, we
have studied in detail the RTPR-catalyzed rate of exchange of
3
3
H from [5′- H]-AdoCbl with solvent in the presence of
Kinetic Parameters Under Conditions in which [AdoCbl]
(9) Booker, S.; Licht, S.; Broderick, J.; Stubbe, J. Biochemistry 1994,
.
[RTPR]. Under standard Michaelis-Menten conditions using
33, 12676-12685.

7466 J. Am. Chem. Soc., Vol. 121, No. 33, 1999
Licht et al.
allosteric effector (dGTP) and in the absence of substrate. We
have shown that this exchange reaction requires homolysis and
reformation of the carbon-cobalt bond and occurs with a rate
-
1
3
constant of 0.3 s (Scheme 1). H2O formation thus places a
lower limit on the rate of hydrogen washout from cofactor, and
its measurement provides information about the timing and
frequency of carbon-cobalt bond reformation. A similar experi-
ment in the presence of substrate ATP, examining the ratio of
3
H2O to dATP as a function of time, could be informative as to
the timing and frequency of carbon-cobalt bond reformation
under these conditions. If most of the release occurs prior to
3
dATP formation, one would expect a burst of H2O. In addition,
the ratio would be expected to decrease as a function of time.
3
If H2O release occurs subsequent to dATP formation one would
expect a burst of dATP and the ratio to increase as a function
of time.
In order to obtain insight into carbon-cobalt bond homolysis
14
3
and reformation, [ C]-ATP and [5′- H]-AdoCbl were incubated
with RTPR, and the reaction was stopped by rapid acid quench
8
14
3
methods. The amounts of [ C]-dATP and H2O were moni-
tored using HPLC and bulb-to-bulb distillation, respectively.
The results are shown in Figure 2A, B, and C. The rate of dATP
-
1
formation (Figure 2A) yields a kapp of 55 ( 10 s . In the first
turnover 0.6 to 0.7 equiv of product are generated.
9
,10
These
results provide the first direct evidence that nucleotide reduction
is not rate-limiting in the steady state.
3
The reaction mixture was also monitored for H2O formation.
-
1
The results are shown in Figure 2B. The kapp is 0.6 s ,
-
1
comparable to the number of 0.3 s previously measured in
the presence of allosteric effector dGTP and in the absence of
substrate. Under the previous conditions of the exchange reaction
3
(
that is, no substrate), H2O release proceeded linearly as a
function of time, in contrast with the behavior shown in Figure
3
2
B. An explanation for the greatly reduced rate of H2O
formation at later times in the presence of substrate is that the
exchange rate is diminished when RTPR is oxidized. These
results are compatible with our previous exchange studies with
4
oxidized RTPR and dGTP.
3
Finally, the ratio of H2O to dATP over time is shown in
Figure 2C. The data are most consistent with H2O release
3
14
3
Figure 2. Formation of [ C]-dATP (A) and [ H]-H
2
O (B) under single
occurring subsequent to dATP formation. This conclusion is
supported by the similarities in the exchange rates in the
presence and absence of substrate. If one C-Co bond cleavage
event effected multiple turnovers of NTPs to dNTPs, then in
the presence of substrates one would expect to observe
substantially reduced rates of 3H2O formation, relative to
experiments carried out in the absence of substrate. Given that
there must be a selection effect on the process of tritium washout
of at least a factor of 10 and a statistical factor of 2 associated
with the 3 hydrogens of 5′-deoxyadenosine, the rate of hydrogen
washout from the cofactor is at least 12 s- and is faster than
turnover conditions. RTPR (120 µM) in one syringe also containing 2
mM [U- C]-ATP, 2 mM dGTP, 20 µM TR, 1 µM TRR, 2 mM
14
NADPH, and 100 mM NaDMG pH 7.3, was mixed with 100 µM [5′-
3
H]-AdoCbl in a second syringe which also contained 2 mM [U-14C]-
ATP, 2 mM dGTP, and 100 mM NaDMG, pH 7.3, followed by
quenching with acid. The reaction mixture was worked up as described
in the Experimental Section. Each data point represents one trial. The
3
data are fitted to a single exponential (solid line). C, The ratio of [ H]-
14
2
H O formed to [ C]-dATP as a function of time, taken from the data
in A and B. The data are fitted to a single exponential.
1
-
1
turnover? This question has now been answered in the case of
the steady-state rate of formation of dATP, at 2 s . These data
thus suggest that C-Co bond reformation occurs faster than
turnover.
RTPR. Results utilizing steady state kinetics with [RTPR] .
[AdoCbl] and results using a presteady state analysis examining
washout of radiolabel from the 5′-position of AdoCbl in the
presence of substrate, both suggest that the chain length of the
reaction is approximately one.
A demonstration that AdoCbl dissociates from RTPR after
each turnover requires carbon-cobalt bond reformation and
would establish a chain length of one. Experiments under steady
state conditions in which [RTPR] . [AdoCbl] were designed
to address this question. Two unusual features of the kinetics
of dNTP formation (measured as disulfide reduction (Scheme
Discussion
In the past five years many protein-based radicals have been
1
identified in enzymes playing primary roles in metabolism.
Their mechanisms are being examined in detail. A mechanistic
question common to all of these reactions is the chain length
of the reaction. Does a single initiation event result in production
of multiple products, or is an initiation event required for each
1
)) under these conditions suggest that turnover is limited by
(10) The inability to generate 1 equiv of dNTP has also been previously
noted. The reasons for this stoichiometry are unclear.
the rate of AdoCbl dissociation. First, the rate does not increase

C-Co Bond Reformation Catalyzed by RTPR
J. Am. Chem. Soc., Vol. 121, No. 33, 1999 7467
3
constant for H2O formation in the presence of substrate is very
similar to the rate constant observed in the absence of substrate.
Furthermore, the rate constant of hydrogen exchange is similar
to the rate constant of dNTP formation in the steady state. The
3
observation that the ratio of H2O release to dATP formation
increases with time during the first turnover suggests that most
3
of the H2O release is associated with reformation of the C-Co
3
bond after dATP formation. Thus, the amount and rate of H
released suggest that carbon-cobalt bond reformation follows
each turnover.
The ability of RTPR to release AdoCbl after each turnover
could have important physiological implications and may
explain the seemingly paradoxical observation that when
Lactobacillus leichmannii cells are grown in the presence of
subsaturating amounts of AdoCbl, the amount of RTPR bio-
Figure 3. Simplified kinetic models for RTPR turnover. In path A,
AdoCbl dissociates from RTPR on each turnover, and in path B AdoCbl
remains bound to RTPR between successive turnovers. E is the reduced
form of RTPR, and F is the oxidized form of RTPR.
11,12
synthesized increases.
Calculations suggest that under these
conditions, L. leichmannii cells produce RTPR in molar excess
over AdoCbl molecules taken up by the cell. A turnover
mechanism involving C-Co bond reformation and AdoCbl
dissociation between turnovers would increase the effective
concentration of holoenzyme, enabling a higher rate of DNA
biosynthesis to be maintained.
linearly with [AdoCbl] as it should if AdoCbl is acting as a
limiting reagent. Instead, in the case of both purine and
pyrimidine nucleotide substrates, saturation kinetics are observed
(Figures 1A and 1B). Second, the turnover numbers Vmax/
[AdoCbl]0 measured under these conditions are larger than the
turnover number Vmax/[RTPR]0 measured under standard Michae-
lis-Menten conditions. The system behaves as if there is more
enzyme turning over substrate than there is AdoCbl available.
This kinetic behavior is consistent with a mechanism in which
AdoCbl dissociates from oxidized RTPR after turnover, binds
to a reduced RTPR and initiates a second turnover faster than
the first molecule of RTPR can be re-reduced.
Experimental Section
Nucleotides, nucleosides, and NADPH were obtained from Sigma.
Alkaline phosphatase (calf intestine) was purchased from Boehringer-
Mannheim. RTPR (specific activity of 1.4 U/mg), TR (specific activity
of 300-700 U/mg), and TRR (specific activity of 3000-7000 U/mg)
were isolated as previously described.1
3-15
C18 Sep-Pak cartridges were
Figure 3 shows a simplified model considering the two
extreme cases: in path A, the AdoCbl dissociates from RTPR
on each turnover, requiring a chain length of one. In path B,
the AdoCbl remains bound under multiple turnovers, and the
chain length would depend upon the frequency with which the
carbon-cobalt bond is reformed. Given the available experi-
mental data, the steady-state assumption, and a few chemically
reasonable assumptions, it is possible to make a crude estimate
of the rate (V) at which dNTPs are generated as a function of
obtained from Whatman. Concentrations of RTPR were measured
-1
-1
spectrophotometrically (ꢀ280 ) 101 000 M cm ), and the enzyme
16
3
was pre-reduced as previously described. [5′- H] AdoCbl was prepared
4
as previously described. All operations involving AdoCbl were carried
out under dim light or red light.
UV-vis spectroscopy was performed on a Cary 3 spectrophotometer
at 37 °C. SF studies were carried out using an Applied Photophysics
DX.17MV stopped-flow spectrophotometer. Data was collected at either
-
1
-1
3
40 nm to monitor consumption of NADPH (∆ꢀ340 ) 6220 M cm ),
or at 525 nm to monitor conversion of AdoCbl to cob(II)alamin (∆ꢀ525
4800 M cm ). A minimum of five traces was used to obtain an
[
AdoCbl] and of the flux through path A: V ) k2(k3 + k5)-
AdoCbl]0/(k2 + k3 + k5) and k5/(k3 + k5), respectively. The
-1
-1
)
[
average kinetic trace. Linear and nonlinear least-squares fits to SF data
were carried out using either the Applied Photophysics system software
or KaleidaGraph. Rapid quench experiments were performed on a Kin-
Tek model RFQ-3 apparatus. To allow reproducible loading in dim
light, the sample loops were loaded using Luer-tip gas-tight syringes
that had been calibrated so that the displacement required to fill each
sample loop was marked on the barrel of the syringe. Reverse phase
HPLC was performed on an Altex HPLC system with an Econosil C18
derivation of this equation is shown in Supporting Information.
The rate constants k1 and k4, binding of NTP and AdoCbl, are
assumed to be fast. The rate constant k2 encompasses many steps
and is known not to be limited by either dNTP formation (55
-1
s , Figure 2A) or cob(II)alamin formation and disappearance
-
1
-1
5,8
(
2
>200 s and 20-50 s , respectively). The value for k2 is
-
1
-1
0-50 s and k3 and/or k6 is 2 s , limited by rereduction of
1
0 µm column (4.6 × 250 mm). The flow rate was 1 mL/min, the
the disulfide and or a conformational change accompanying this
process.5 The rate constant k5 is 8-10 s-1 as measured in
Figure 1A and B. The reduced form of RTPR is designated E,
and the oxidized form is designated F. With this information,
it is clear that the flux through path A predominates (k5/(k3 +
k5) ) 0.8). The dissociation of AdoCbl on each turnover requires
that the carbon-cobalt bond is reformed on every turnover,
establishing a chain length for the reaction of approximately
one. This simplified model also predicts quite well the rate of
dNTP formation as a function of [AdoCbl] (data not shown).
elution profile was monitored by A260, and fractions of 1 mL were
collected. Scintillation counting was performed on a Beckman LS 6500
scintillation counter using Scint-A scintillation fluid at a ratio of 8.5
mL Scint-A per mL of eluate.
,8
[AdoCbl]-Dependence of the Rate of Turnover when [RTPR]
. [AdoCbl]: Measurement of the Rate of AdoCbl Dissociation from
RTPR. Assays contained in a volume of 500 µL: 200 mM sodium
dimethylglutarate (NaDMG) pH 7.3, 1 mM ATP, 1 mM dGTP, 0.25
mM NADPH, AdoCbl (0.1-0.8 µM), 85 µM TR, 1.5 µM TRR, and
RTPR (1-8 µM). RTPR was present in 10-fold excess over AdoCbl.
Further support for a chain length of one is obtained from
H2O release experiments using [5′- H]-AdoCbl under turnover
(11) Davis, R. L.; Layton, L. L.; Chow, B. F. Proc. Soc. Exptl. Biol.
Med. 1952, 79, 273-276.
3
3
(
12) Goulian, M.; Beck, W. S. J. Biol. Chem. 1966, 241, 4233-4242.
13) Booker, S.; Stubbe, J. Proc. Natl. Acad. Sci. U.S.A. 1993, 90, 8352-
conditions. If a single C-Co bond cleavage event resulted in
(
formation of multiple dNTPs, then one would expect the amount
8356.
3
of H2O released to be very low relative to the amount of dNTP
(14) Lunn, C. A.; Kathju, S.; Wallace, C.; Kushner, S.; Pigiet, V. J. Biol.
Chem. 1984, 259, 10469-10474.
generated, since the carbon-cobalt bond is not reformed and
(
15) Russel, M.; Model, P. J. Bacteriol. 1985, 163, 238-242.
3
hence there is no opportunity for washout of H from the
(16) Booker, S. Ph.D. Thesis. Massachusetts Institute of Technology,
cofactor. Instead, what is observed is that the apparent rate
1994.

7468 J. Am. Chem. Soc., Vol. 121, No. 33, 1999
Licht et al.
All components except AdoCbl were mixed in a 500 µL quartz cuvette
and incubated at 37 °C for 1-2 min. A background rate of NADPH
consumption was measured by monitoring A340. AdoCbl was added to
the cuvette, and the rate of NADPH consumption was measured. Similar
experiments were carried out using 1 mM CTP and 1 mM dATP. Under
the conditions above where [AdoCbl] > 0.8 µM, SF UV-vis methods
were required to measure the rate of NADPH consumption. RTPR (10-
which was immediately capped. The flask containing the distillate was
washed with 500 µL of water, which was also transferred to the
scintillation vial. Scintillation fluid (8.5 mL) was added to the vial,
and the distillate was counted.
The lyophilized material was redissolved in 1 mL of water, frozen
on dry ice, and stored at -20 °C. It was then thawed and loaded onto
reverse phase C18 Sep-Paks. Nucleotides were eluted with 10 mL of
3
0 µM), 1 mM ATP (1 mM CTP), 1 mM dGTP (1 mM dATP), 175
µM TR, 3 µM TRR, and 0.5 mM NADPH in 200 mM NaDMG, pH
.3, were placed in one syringe and mixed with an equal volume from
3
water. Nucleosides and AdoCbl were eluted with 50% (v/v) CH OH
(aqueous) (10 mL), and the solvent was removed by lyophilization.
The nucleotide-containing samples were lyophilized, then redissolved
in 50 mM Tris, 0.1 mM EDTA (970 µL) to which 30 U of alkaline
phosphatase was added. The reaction was incubated at 37 °C for 3 h.
The reaction mixture was analyzed by HPLC with a linear gradient.
7
a second syringe of the same reaction buffer containing 1-3 µM
AdoCbl, 1 mM ATP (1 mM CTP), 1 mM dGTP (1 mM dATP), and
consumption of NADPH was monitored. Experiments were also carried
out with [RTPR] < [AdoCbl]. Assays contained in a volume of 500
µL: NaDMG pH 7.3, 1 mM ATP (1 mM CTP), 1 mM dGTP (1 mM
dATP), 0.25 mM NADPH, AdoCbl (0.6-4 µM), and RTPR (0.1-
Solvent A was H O and solvent B was MeOH: 0-2 min, isocratic
2
elution with A; 2-7 min, 0-20% B; 7-30 min, 20% B. Adenine,
adenosine, and 2′-deoxyadenosine (2′-dA) eluted at 12, 16.5, and 18
min, respectively. Fractions containing adenosine and 2′-dA were pooled
and lyophilized. Because this HPLC method effected only partial
separation of adenosine and 2′-dA, the 2′-dA containing fractions were
further chromatographed by the method of Cory et al.17 Borate columns
(Dowex AG 1 × 2, exchanged into the borate form using sodium
tetraborate) were poured into glass wool-stoppered 5.75 in. Pasteur
pipettes, washed with 20 mL of water, and then equilibrated by washing
with 1 mM sodium tetraborate (2 mL). The lyophilized samples were
redissolved in 1 mL of water and loaded onto the borate columns. The
columns were washed with 2 mL of 1 mM sodium tetraborate, and
2′-dA was eluted with 12 mL of 1 mM sodium tetraborate. UV-vis
0
.75 µM). AdoCbl was present in 5-fold excess over RTPR.
Rapid Acid Quench Experiments Under Single Turnover Condi-
3
tions: Fate of the 5′-Hydrogens of [5′- H]-AdoCbl and Rate of dATP
Formation. Pre-reduced RTPR (120 µM), 20 µM TR, 1 µM TRR, 2
14
6
mM NADPH, 2 mM [U- C]-ATP (5.2 × 10 cpm/µmol), 2 mM dGTP,
and 100 mM NaDMG pH 7.3, were placed in one syringe and rapidly
mixed at 37 °C with an equal volume from a second syringe of [5′-
3
6
H]-AdoCbl (100 µM, 1.2 × 10 cpm/µmol), 2 mM ATP, and 2 mM
dGTP in the same reaction buffer. After the specified time, the reaction
was quenched with 2% (v/v) perchloric acid (60-220 µL), collected
in tubes containing 5′-dA (55 nmol) and 5′,8-cycloadenosine (15 nmol).
Immediately after being quenched, samples were neutralized with equal
volumes of 0.4 M KOH and 0.5 M NaDMG pH 7.3 (50-200 µL each).
Neutralized samples were immediately quick-frozen in liquid nitrogen
and stored on dry ice. The neutralized reaction mixtures were stable at
-1
-1
spectroscopy (ꢀ260 ) 15 200 M cm for 2′-dA) was used to measure
the recovery. The sample was lyophilized and redissolved in 1 mL
water, and radioactivity was analyzed by scintillation counting.
Acknowledgment. This research was supported by a grant
from the National Institutes of Health (GM 29595) to J.S. S.L.
was a recipient of a Howard Hughes predoctoral fellowship
award.
-
20 °C for up to a week. A zero time point was generated by omitting
5′- H]-AdoCbl from the second syringe and putting it instead in the
3
[
sample collection tube, so that the protein would be acid-precipitated
before encountering it. Tritiated water was removed from the quenched
samples by bulb-to-bulb distillation. All operations were carried out in
a fume hood. Samples (∼1 mL) were transferred to silanized 50-mL
pear-shaped flasks and shell-frozen in dry ice/acetone. The bulb-to-
bulb distillation apparatus (two pear-shaped flasks connected by a
Y-shaped adapter equipped with a stopcock) was evacuated and kept
under static vacuum. The flask containing the sample was removed
from the dry ice/acetone bath, and the empty flask was placed in the
dry ice/acetone bath. Lyophilization of the sample was usually complete
in 45 min or less. The distillate was transferred to a scintillation vial,
Supporting Information Available: Derivation of the rate
equation for dNTP formation according to the model in Figure
3 (PDF). This material is available free of charge via the Internet
at http://pubs.acs.org.
JA9913840
(17) Cory, J. G.; Russell, F. A.; Mansell, M. M. Anal. Biochem. 1973,
55, 449-456.