Tài liệu Báo cáo khoa học: Kinetics and thermodynamics of nick sealing by T4 DNA ligase pptx

Kinetics and thermodynamics of nick sealing by T4 DNA ligase
Alexey V. Cherepanov* and Simon de Vries
Kluyver Department of Biotechnology, Delft University of Technology, the Netherlands
T4 DNA ligase is an Mg
2+
-dependent and ATP-dependent
enzyme that seals DNA nicks in three steps: it covalently
binds AMP, transadenylates the nick phosphate, and cata-
lyses formation of the phosphodiester bond releasing AMP.
In this kinetic study, we further detail the reaction mechan-
ism, showing that the overall ligation reaction is a super-
imposition of two parallel processes: a ễprocessiveế ligation, in
which the enzyme transadenylates and seals the nick without
dissociating from dsDNA, and a ễnonprocessiveế ligation, in
which the enzyme takes part in the abortive adenylation
cycle (covalent binding of AMP, transadenylation of the
nick, and dissociation). At low concentrations of ATP
(< 10 l
M
) and when the DNA nick is sealed with
mismatching base pairs (e.g. ve adjacent), this super-
imposition resolves into two kinetic phases, a burst
ligation (% 0.2 min
)1
) and a subsequent slow ligation
(% 2 ã 10
)3
min
)1
). The relative rate and extent of each
phase depend on the concentrations of ATP and Mg

2+
.The
activation energies of self-adenylation (16.2 kcalặmol
)1
),
transadenylation of the nick (0.9 kcalặmol
)1
), and nick-
sealing (16.318.8 kcalặmol
)1
) were determined for several
DNA substrates. The low activation energy of transadeny-
lation implies that the transfer of AMP to the terminal DNA
phosphate is a spontaneous reaction, and that the T4 DNA
ligaseAMP complex is a high-energy intermediate. To
summarize current ndings in the DNA ligation eld, we
delineate a kinetic mechanism of T4 DNA ligase catalysis.
Keywords: DNA ligase; end-joining; kinetics; mechanism of
action; mismatching nick.
T4 DNA ligase is an enzyme that catalyses formation of the
phosphodiester bond between the adjacent 5Â-PO
4
and
3Â-OH groups of two dsDNA fragments [1]. It is able to join
two dsDNAs (blunt-end or sticky-end ligation), or seal a
break between two ssDNA fragments annealed on the
complementary DNA strand (nick-ligation). It can join
phosphodiester linkages on triple-stranded nucleic acids [2],
seal single-stranded 15-nucleotide gaps [3], and act as a
lyase, removing apurininc/apyrimidinic (AP) sites in DNA

[4]. The enzyme requires a bivalent metal cation such as
Mg
2+
or Mn
2+
, and joins DNA using ATP as a coenzyme.
One phosphodiester bond in DNA is formed per ATP
molecule hydrolysed to AMP and pyrophosphate.
The mechanism of catalysis of T4 DNA ligase comprises
three steps and involves two covalent reaction intermedi-
ates:
E ỵ ATP $ EAMP ỵ PP
i
1ị
E AMP ỵ ndsDNA $ EAMPndsDNA
! E ỵ AMPndsDNA 2ị
E ỵ AMP ndsDNA $ E AMPndsDNA
! E ỵ dsDNA ỵ AMP 3ị
where ndsDNA is nicked dsDNA, AMPndsDNA is
ndsDNA adenylated at the 5Â-phosphate of the nick, and
a one-sided arrow indicates that the ễreverseế reaction is
at least three orders of magnitude slower than the
ễforwardế reaction.
On the basis of the electrophoretic mobility shift assay
experiments, it has been suggested that adenylated ligase
forms transient Tcomplexes, EAMPndsDNA, in search
of a phosphorylated 5Â-end of (n)dsDNA. When the nick
phosphate is found, it is adenylated, and a stable ễScomplexế
is formed, EAMPndsDNA [5]. The enzyme in this
complex has been suggested to ễstallế on DNA until the

nick is sealed and dsDNA is released. Formation of the rst
phosphodiester bond during joining of the blunt ends has
been proposed to happen accordingly; the main difference is
the 2 : 1 dsDNA to enzyme stoichiometry in step 3. It has
been suggested [5] that during blunt end joining, the ternary
complex ligaseDNA is formed via two second-order
associative reactions:
E ỵ AMPdsDNA ! EAMPsDNA (Scomplex)
Scomplex ỵ dsDNA ! EAMPdDNAdsDNA
Despite a good understanding of the overall reaction
mechanism, relatively few articles have been dedicated
to the kinetic studies of catalysis performed by this
enzyme. The optimal concentration of Mg
2+
in the nick-
joining reaction (810 m
M
), apparent K
m
for the nicked
Correspondence to S. de Vries, Kluyver Department of Biotechno-
logy, Delft University of Technology, Julianalaan 67,
2628 BC Delft, the Netherlands.
Fax: + 31 15 2782355, Tel.: + 31 15 2785139,
E-mail:
Abbreviations: ndsDNA, nicked dsDNA.
Enzymes: DNA ligase (EC 6.5.1.1).
*Present address: Metalloprotein & Protein Engineering Group,
Leiden Institute of Chemistry, Gorlaeus Laboratories, Leiden
University, Einsteinweg 55, PO Box 9502, 2300 RA Leiden,

the Netherlands.
(Received 28 May 2003, revised 8 September 2003,
accepted 9 September 2003)
Eur. J. Biochem. 270, 43154325 (2003) ể FEBS 2003 doi:10.1046/j.1432-1033.2003.03824.x
dsDNA (1.5 n
M
), and ATP (14–20 l
M
) as well as the
apparent inhibition constant K
i
for dATP (10–35 l
M
)
had been determined during the initial characterization
of the enzyme [6,7]. It was shown that T4 DNA ligase
binds ATP covalently forming a lysine (e-amino)-linked
adenosine monophosphoramidate [8]. Harvey et al. [9]
have demonstrated that sealing of the pre-adenylated
nick in dsDNA is inhibited when T4 DNA ligase is
preincubated with ATP and Mg
2+
. Later it was shown
that the enzyme obeys Ping-Pong kinetics, and joins
dsDNA with multiple nicks in the nonprocessive mode.
The true K
m
for ATP (100 l
M
) was determined in the

joining reaction with polydAÁd(pT)
10
as substrate (true
K
m
¼ 0.6 l
M
) [10]. It was shown that T4 DNA ligase
seals nicks containing base pair mismatches [11–15], and
the effect of the ionic strength on the apparent K
m
of the
mismatching dsDNA nick has been evaluated (e.g.
200 n
M
at 0.2
M
NaCl vs. 50 n
M
without salt)
[12,16,17]. The pH-dependent equilibrium constant for
step 1 (0.0213 at pH ¼ 7.0 and 25 °C, pK
a
¼ 8.4) and
the standard free energy for cleavage of ATP to AMP
()10.9 kcalÆmol
)1
) have been determined [18]. It was
shown that T4 DNA ligase is capable of synthesizing the
dinucleoside polyphosphates, such as Ap

3
A, Ap
4
A,
Ap
4
G, and Ap
4
dA, using ADP (d)ATP, GTP, and P
3
as substrates [19,20]. This secondary enzyme activity
may stem from the fact that ligase has two closely
located nucleotide-binding sites [21]. Recently, a
dynamic mechanism of nick recognition by DNA ligase
has been proposed [22]. The key feature of the
mechanism is the B-to-A DNA helix transition of the
enzyme-bound dsDNA motif, which results in DNA
contraction, bending, and unwinding. For non-nicked
dsDNA, this transition is reversible, leading to dissoci-
ation of the enzyme. For ndsDNA, this transition was
proposed to (a) trigger an opened–closed conforma-
tional change in the enzyme, and (b) force the motif to
accommodate the strained A/B-form hybrid conforma-
tion, the transition state in the nick-sealing reaction.
In our previous work, we assessed the ability of T4 DNA
ligase to seal ndsDNA containing one to five adjacent
mismatching base pairs [14,23], aiming to use this enzyme in
the novel protocol of saturated scanning mutagenesis. In all
cases, kinetic traces displayed pronounced biphasic beha-
vior, which was most spectacular with the nick containing

five base pair mismatches. Apparently, this effect has not
been previously reported. To understand its origin, we
decided to study the mechanism of T4 DNA ligase in more
detail. We have previously reported a pre-steady-state
kinetic analysis of the first step of DNA ligase catalysis
(covalent binding of AMP [24]), showing that the enzyme
employs a two-metal-ion mechanism for this nucleotidyl
transfer reaction, using the dimagnesium ATP form
(ATPÁMg
2
) as a true substrate. The monocoordinated
form, ATPÁMg, and/or free ATP bind DNA ligase
noncovalently, with K
d
< 150 n
M
[21]. Nucleotidyl transfer
is reversible, and the monomagnesium pyrophosphate form
MgÁP
2
O
7
participates in the T4 DNA ligase-promoted
synthesis of ATP [24].
This work concentrates on the steady-state kinetic
analysis of the overall ligation reaction and an initial
thermodynamic characterization of the catalysis. We aimed
to achieve the following goals: (a) to deepen the general
understanding of the kinetic mechanism of the end-joining
reaction, the biphasic behavior in particular, using an

ndsDNA substrate with 5¢-mismatching base pairs; (b) to
extract kinetic and thermodynamic parameters of the
elementary steps of ligase catalysis; and (c) to summarize
current findings on the mechanism of action of DNA ligase
in a single reaction scheme.
Experimental procedures
Enzymes and oligonucleotides
Three commercial batches of T4 DNA ligase were used, and
were purchased from Amersham Biosciences (Uppsala,
Sweden), Roche Molecular Biochemicals (Basel, Switzer-
land), and MBI Fermentas (Vilnius, Lithuania). All enzyme
batches showed similar activity and substrate specificity.
The proteins were essentially pure as judged by SDS/PAGE.
Protein concentration in the purchased enzyme stocks was
determined using the BCA protein determination kit (Pierce
Biotechnology, Rockford, IL, USA). Synthetic oligonucleo-
tides were purchased with Eurogentec (Seraing, Belgium),
except for the one labeled with the Cy5 fluorescent marker,
which was obtained from Amersham Biosciences.
Model system
For studies on T4 DNA ligase-promoted repair of nicks in
dsDNA, we used 72/24/(6–24)-mer synthetic DNA sub-
strates. Nonphosphorylated 72-mer B had the sequence
5¢-GTCCAAACAGCTATCTGCATCCGTCGACCTGC
TCGGTTCCTTGGCTACACTGGCCGTCGTTTTACA
ACGTCG-3¢.The24-mer5¢-DNA oligonucleotide C
(5¢- here refers to the fact that the 5¢-oligomer is located
upstream of the nick) had the sequence 5¢-CGACGTT
GTAAAACGACGGCCAGT-3¢, and contained on the
5¢-end the fluorescent label Cy5 (Dye 667, No. 27-1801-02;

Amersham Biosciences), allowing easy quantification of the
products of the joining reaction. 3¢-DNA oligonucleotides
(located downstream of the joining site) of different lengths
(6–24-mers; Figs 1, 2, 3, 4 and 8) were 5¢-phosphorylated;
oligonucleotides M5C19, M1C6, and M4C7, in addition,
contained base pair mismatches next to the joining site.
Ligation of ndsDNA
Theligationreactionwasperformedin30lL66m
M
Tris/
HCl/1 m
M
dithiothreitol/0.05 mgÆmL
)1
BSA, pH ¼ 7.6.
The buffer pH (measured at +20 °C) was adjusted to 7.19
(7.34) when the ligation was performed at +4 °C(+10°C),
counting dpH/°C ¼ )0.026 for the Tris/HCl buffer pair. To
follow the formation of the adenylated DNA intermediate,
[
32
P]ATP[aP] was used. Concentrations of ATP, MgCl
2
,
oligonucleotides, and the incubation temperature were
varied according to the comments in the text. ndsDNA
was prepared by mixing the necessary amount of 72-mer
oligonucleotide B, a stoichiometric amount of 5¢-Cy5-
labeled 24-mer C, and the required donor oligonucleotide,
followed by 5 min incubation at 65 °C, 5 min incubation at

37 °C, and 10 min incubation at room temperature. T4
DNA ligation buffer was added after pre-annealing of the
4316 A. V. Cherepanov and S. de Vries (Eur. J. Biochem. 270) Ó FEBS 2003
oligonucleotide and before addition of the enzyme; the
reaction mixture was preincubated at the assay temperature
for 5 min. The reaction was initiated by addition of the
enzyme. Over time, aliquots of 0.5 lL were withdrawn and
mixedwith10lL 100% formamide/10 m
M
NaOH/10 m
M
EDTA/5 mgÆmL
)1
Blue Dextran, pH ¼ 9.5.
Stability of DNA ligase at ambient temperature
It is known that preparations of T4 DNA (RNA) ligase
gradually lose their activity when incubated at temperatures
above 0 °C [6,25]. In this work, ligase was assayed at +4 °C
for periods up to 25–40 h. To avoid irreversible inactivation
of T4 DNA ligase during prolonged incubations, BSA was
added to the reaction mixture to a concentration of
0.05 mgÆmL
)1
, as reported previously [21].
Separation of ligation products and data analysis
All DNA separations were performed on an ALF Express
DNA sequencer (Amersham Biosciences) using 6–15%
acrylamide gels (Tris/borate/EDTA/7
M
urea). Usually,

runs were performed at 55 °C, with 80 mA current and
lasted for 1–3 h. For each data point, the fluorescence of
two chromatographically separated peaks of starting mater-
ial and ligation product was obtained. Separation of
[
32
P]DNA was visualized using a Molecular Dynamics
Phosphorimager SI (Amersham Biosciences) and quantified
using
IMAGEQUANT
software. Data obtained from both
Cy5-labeled and
32
P-labeled DNA was imported into
IGOR
PRO
version 4.0 (WaveMetrics Inc., Salt Lake City, UT,
USA); further data analysis such as integration of peaks and
fitting was performed using the built-in functions of this
software package. Rate values for the burst ligation were
determined as V
init
¼ F ¢[t]
t ¼ 0
, where F [t] is the exponential
fitting function. Steady-state rates were determined by
interpolating the steady-state region of the product forma-
tion curve with a linear regression function. Numbers of
turnovers were calculated by dividing the initial concentra-
tion of ndsDNA in the reaction mixture by the concentra-

tion of T4 DNA ligase.
Pre-steady-state kinetic analysis
Transient-state kinetic experiments were performed on the
Bio Sequential Stopped-Flow Reaction Analyzer SX-18MV
(Applied Photophysics, Leatherhead, Surrey, UK) using the
ozone-free 150 W xenon-arc light source. The
SX
-18
MV
software package for a single-wavelength operation
mode was used for the optical measurements. Tryptophan
emission was excited at 280 nm and measured as the light
passing through a < 320 nm cut-off filter. Kinetic traces of
protein fluorescence emission were obtained by averaging
three to ten shots. Error estimates for the data values in
graphs and tables represent 95% confidence intervals
calculated using Student’s distribution function.
In the stopped-flow instrument, we studied the transient-
state kinetics of the binding of Mg
2+
to the EÁATP complex
at different pH values. For this experiment, a ligase solution
(7.5 l
M
, 0.41 mgÆmL
)1
) was prepared by dilution of
a10mgÆmL
)1
enzyme stock into 0.075 mgÆmL

)1
BSA
solution in deionized water containing 1.5 m
M
dithioerythr-
itol and % 230 l
M
ATP(ATPwasaddedfroma50-m
M
stock solution pre-equilibrated to pH ¼ 7). This weakly
buffered solution at pH % 7.5 was stored on ice until
further use. A 150 m
M
Tris/HCl buffer of the desired pH
was prepared separately, as well as the 15 m
M
solution of
MgCl
2
in deionized water. At 5 min before the mixing shots,
150 m
M
Tris/HCl buffer was diluted threefold into both
enzyme and Mg
2+
stocks. Then 200-lL aliquots of the
resulting solutions were withdrawn, mixed with each other,
and the pH of the mixture measured. In parallel, the
solutions were pre-equilibrated to ambient temperature in
the drive syringes and rapidly mixed in the stopped flow

instrument, triggering the reaction.
Results and discussion
Time course of the joining reaction
To study the T4 DNA ligase-promoted end-joining
reaction, we used synthetic DNA oligonucleotides as
described in Experimental procedures. Two ndsDNA
substrates were assembled: a substrate containing a 5-bp
mismatching fragment at the nick (BÁMÁ5C19), and a
complementary nick substrate, BÁC24. T4 DNA ligase
effectively utilizes both of these substrates (Figs 1, 3 and 4).
Fig. 1. Biphasic kinetics of joining of 3¢-oligonucleotides (M5C19, C24,
M1C6, and M4C7) to 72/24-mer BÁC. [dsDNA] was 1 l
M
;[ATP]was
5.6 l
M
for M5C19, C24, and 1 m
M
for M1C6 and M4C7. Ligation of
M5C19 and C24 was performed at +10 °C as described in Experi-
mental procedures. M1C6 and M4C7 were joined at +4 °Caspre-
viously described [14]. For joining of C24, the concentration of T4
DNA ligase in the assay mixture was 8 n
M
, 0.1 l
M
for M5C19, and
0.4 l
M
for M1C6 and M4C7. [dsDNA] (1 l

M
) corresponds to the 125
turnovers of T4 DNA ligase in the case of C24 joining, 10 turnovers in
the case of M5C19, and 2.5 turnovers for M1C6 and M4C7.
Ó FEBS 2003 Kinetics of nick sealing by T4 DNA ligase (Eur. J. Biochem. 270) 4317
The mismatching nick (BÁMÁ5C19) is sealed in two kinetic
phases (Fig. 1A). In the case of C24, these phases are less
pronounced, and are observed only at low concentrations
of ATP (Fig. 1B). Similar biphasic behavior is observed
with nicks containing one to four mismatching base pairs
(e.g. Fig. 1C,D). The origin of the two phases becomes
clear when the formation trace of the adenylated ndsDNA
intermediate is plotted together with the trace of the
ligation product (Fig. 2). Virtually all ndsDNA is conver-
ted into AMP-dsDNA before 20% of the ligation product
is formed, and during this period the initial burst phase of
ligation takes place. The slow ligation phase starts when
all available DNA substrate and T4 DNA ligase are
converted into their respective AMP-bound intermediates.
The following mechanistic interpretation of the biphasic
kinetics is suggested. During the burst phase, the enzyme
performs ligation ÔprocessivelyÕ, i.e. by transadenylating
and sealing the nick without dissociating from the DNA
complex or being re-adenylated. This process is not ÔidealÕ:
a fraction of ligase molecules dissociates after the transfer
of AMP to the nick phosphate, rebinds AMP, and
performs another transadenylation step, which in parallel
to the ÔprocessiveÕ ligation leads to the accumulation of
AMP–ndsDNA. The slow steady-state ligation phase
starts when the concentration of the AMP–ndsDNA

intermediate reaches its maximum. The overall end-joining
rate decreases because the adenylated DNA is ligated
either by the adenylated enzyme with a notably lower
rate, or, and what is more likely, by a small fraction of
the AMP-free enzyme (in agreement with previous data
[9,14]). It is also clear that the formation of the
phosphodiester bond rate-limits sealing of the mismatch-
ing nicks, and not the adenylation of the 5¢-nick
phosphate. For example, the rate of burst ligation of
the oligonucleotide M1C6 (19 h
)1
) is, within experimental
error, identical with the sealing rate of pre-adenylated
AMP–M1C6 (18 h
)1
) [14]; adenylation of the oligonucle-
otide M5C19 (1 min
)1
) is almost fivefold faster than the
burst ligation (0.2 min
)1
).
A complex of T4 DNA ligase with AMP–ndsDNA is
more stable in the case of a complementary nick ([5]), and
the enzyme hardly dissociates from ndsDNA between the
steps of transadenylation and nick-sealing. As a result, the
steady-state concentration of AMP–ndsDNA during liga-
tion would be lower, and a difference in joining rates
between the burst ligation and slow ligation phases at the
same [ATP] is less pronounced. For example, in the case of

BÁCÁ24, the rate of the burst phase is only about sixfold
higher than the rate of the slow phase, in contrast with
> 100-fold difference in the case of the mismatching nick
(Fig. 1). The amplitude of the burst phase reaches % 50% of
the total extent of ligation when [ATP] is taken of the same
Fig. 2. Formation of the ligation product (d) and the AMP–DNA
intermediate (s) during the joining of M5C19–72/24-mer BÆC. Ligation
was performed at +10 °C. T4 DNA ligase (0.4 l
M
)sealedndsDNA
(1 l
M
) in the presence of 1 m
M
ATP and 5 m
M
MgCl
2
.Dottedtraces
were obtained by fitting exponential functions to the experimental data
points. [TP]
0
represents the number of turnovers required for the
joining of all ndsDNA in the reaction mixture.
Fig. 3. Joining of C24 to 72/24-mer BÆA at different concentrations of
ATP. (A) Product formation curves were computed by fitting single/
double exponential functions to the experimental data points. [ATP]
0
for each curve is shown in the inset. The linear/steady-state phase of
the joining reaction is magnified in the inset. Rate values were

determined by fitting linear regression to the kinetic traces (fitted
traces are shown in the inset). (B) (d) Turnover for the ligation of
C24 and (s) M5C19 (burst ligation) at different [ATP]. The 0–50 l
M
region of ATP concentrations is magnified in the inset. (C) Line-
weaver–Burk plot of the (d) data shown in (B). Ligation was per-
formed under the conditions described in Fig. 1. The concentration
of Mg
2+
was 5.1 m
M
.
4318 A. V. Cherepanov and S. de Vries (Eur. J. Biochem. 270) Ó FEBS 2003
order of magnitude as [ndsDNA] in the reaction mixture,
e.g. [ATP] % 5 l
M
(Figs 2 and 4).
In general, biphasic kinetics are observed because the
ligase in complex with AMP–ndsDNA does not re-adeny-
late itself. Otherwise, the enzyme would already be saturated
with AMP at the beginning of the reaction, irrespective of
the presence of ndsDNA, and would promote only slow
ligation. In this sense, AMP–ndsDNA seems to ÔshieldÕ
ligase from ATP. Biphasic kinetics support the previous
proposals that: (a) the adenylyl moiety in the ligaseÁAMP–
ndsDNA complex occupies the ATP-binding pocket of the
protein [5], preventing a second nucleotide molecule from
entering the active site; and/or (b) ligase-bound ndsDNA
covers the ATP-binding site, hindering the access of ATP
from solution [26,27].

Effect of ATP concentration
The rate of the end-joining reaction catalysed by T4 DNA
ligase depends on the concentration of ATP in the reaction
mixture. For example, the sealing rate of the complementary
nick in BÁCÁ24 increases more than 10-fold with increasing
[ATP] from 5 to 400 l
M
, yielding K
app
m
(ATP) ¼ 1.1 ·
10
)4
M
(Fig. 3), similar to the previously reported value
obtained with poly(dA)Áoligo(dT)
10
[10]. There are two
possible explanations for the decrease in the end-joining rate
at low [ATP]: either the binding of the nucleotide to the
ligase becomes rate limiting, or the equilibrium shifts
towards the nonadenylated enzyme. From the pre-steady-
state kinetic analysis of ATP binding to T4 DNA ligase in
the absence of ndsDNA [24], both of these possibilities seem
unlikely. At 5 m
M
Mg
2+
, the rate constant for binding of
ATP is k

on
% 9 · 10
5
M
)1
Æs
)1
, giving a binding rate of
4.5 s
)1
at 5 l
M
ATP, which is more than 100-fold faster
than the observed rate of ligation at this [ATP] (% 1min
)1
).
Therefore, binding of ATP could by no means limit the
enzyme turnover, unless inhibition by DNA is considered.
Furthermore, the apparent K
d
for ATP in the noncovalent
EÁATP complex is below 150 n
M
[21], implying that at 5 l
M
ATP the concentration of free enzyme in the absence of
ndsDNA is negligible; the ligase is essentially ATP bound
and/or AMP bound.
Pre-steady-state kinetic experiments in which binding of
(n)dsDNA to T4 DNA ligase was studied suggest that this

is a rapid process with k
obs
% 10
8
M
)1
Æs
)1
(A. Cherepanov,
D. Pyshny & V. Chikaev, unpublished results). Thus,
the binding of DNA at low [ATP] would be notably faster
than binding of ATP (10
2
s
)1
for dsDNA vs. 4.5 s
)1
for
ATP at 1 l
M
ndsDNA and 5 l
M
ATP). The situation is
reversed at high ATP concentrations, when binding of ATP
is faster than binding of DNA (10
2
s
)1
for DNA vs.
4.5 · 10

3
s
)1
for ATP at 1 l
M
ndsDNA and 5 m
M
ATP).
Modeling studies indicate that ndsDNA in complex with
ligase hinders the access of solution ATP to the nucleotide-
binding site [26,27]. Therefore, it is likely that the decrease in
the nick-sealing rate at low [ATP] occurs because DNA
prevents binding of ATP to the ligase. This agrees with the
fact that the joining rate of the mismatching oligonucleotide
M5C19 does not decrease with [ATP] in the range 5 l
M
to
3m
M
(Fig. 4). In contrast, an approximately twofold
increase in the joining rate is observed, compared with
more than 10-fold decrease in the latter in the case of C24.
In the case of M5C19, formation of the phosphodiester
bond (< 0.2 min
)1
) limits the rate of the enzyme turnover,
being slower than binding of a DNA substrate (10
2
s
)1

),
ATP (> 4.5 s
)1
), and joining of the complementary nick
under the same conditions (> 1 min
)1
, Table 1). The
increase in the joining rate of M5C19 at low [ATP] may
indicate a shift of equilibrium towards the catalytically
competent nonadenylated form of the enzyme.
At [ATP] above 3 m
M
, inhibition of joining is observed:
at 10 m
M
ATP the nick-sealing rate is roughly 20-fold lower
than at 1 m
M
ATP (Fig. 3). T4 DNA ligase is known to
synthesize dinucleoside polyphosphates, such as Ap
4
A[19].
In this reaction the enzyme binds a second ATP molecule in
the DNA-binding site with K
d
0.1–0.25 m
M
[21]; the Ap
4
A

synthesis is inhibited by ndsDNA [20]. In our case, at
low [ndsDNA] and high [ATP], the opposite situation
may arise, when ATP inhibits binding of ndsDNA and
subsequent ligation by occupying the dsDNA-binding site,
leading to the synthesis of Ap
4
A (similar to that reported for
GTP [20]).
Fig. 4. Joining of M5C19 to 72/24-mer BÆC at different concentrations
of ATP. (A) Product formation curves at T ¼ 10 °C. Concentration of
ATP is shown in the graph. (h)130l
M
ATP; (j)362l
M
;(s)
1.16 m
M
;(d)3.15m
M
. The burst-ligation phase of joining is magni-
fied in the inset. Reaction turnovers of the burst ligation (C) and of the
slow ligation (D) are weakly [ATP]-dependent, increasing at
[ATP] < 100 l
M
% 2-fold. The amplitude of the burst-ligation phase
(B) increases more than 10-fold from % 0.3 turnovers at 3.15 m
M
ATP
to 4.5 turnovers at 5.5 l
M

ATP, and the increase starts at
[ATP] < 100 l
M
. Ligation was performed under conditions equival-
ent to those described in Fig. 1. The concentration of Mg
2+
was 5 m
M
.
[dsDNA] (1 l
M
) corresponds to 10 enzyme turnovers (T4 DNA ligase
was 0.1 l
M
).
Ó FEBS 2003 Kinetics of nick sealing by T4 DNA ligase (Eur. J. Biochem. 270) 4319
In the case of joining of the mismatching oligonucleotide
M5C19, the amplitude of the burst phase increases more
than 10-fold from % 0.3 turnovers at 3 m
M
ATP to 4.5
turnovers at 5 l
M
ATP (Fig. 4), while the joining rate
remains roughly the same. The amplitude of the slow phase
(extent of ligation) is independent of the concentration of
ATP, approaching % 90% of the total concentration of the
ndsDNA substrate in the reaction mixture.
Effect of Mg
2+

concentration
Mg
2+
is essential for DNA joining catalysed by T4 DNA
ligase. It has previously been shown that the optimal
concentration of Mg
2+
in the ligation mixture is % 10 m
M
[6]. As shown in Fig. 5A, two kinetic phases of the end-
joining of M5C19 have different requirements for [Mg
2+
].
The rate of burst ligation increases with [Mg
2+
], and the
optimum is above 5 m
M
Mg
2+
. In contrast, the rate of
the slow phase of joining reaches its maximum at % 1m
M
Mg
2+
, above which joining is inhibited. It is interesting
that the dependence of joining rates on [Mg
2+
] follows the
changes in the equilibrium concentrations of the mono-

magnesium and dimagnesium forms of ATP (Fig. 5B).
The maximal rate of the slow phase of the joining of
the mismatching oligonucleotide M5C19 corresponds to
the maximal concentration of ATPÁMg. On the other
hand, the rate of the burst phase of joining of M5C19,
and the joining rate of the complementary oligonucleotide
C24 resembles more the increase in the concentration of
ATPÁMg
2
.
Temperature-dependence and pH-dependence
of self-adenylation of T4 DNA ligase
We have previously shown that self-adenylation of T4 DNA
ligase in the absence of ndsDNA proceeds according to
Scheme 1 [24].
Fig. 5. [Mg
2+
]-dependence of the joining of M5C19 (A) or C24 (inset in
A) to 72/24-mer BÆC and the equilibrium concentrations of different ATP
forms (B). (A) Turnovers of the burst ligation for M5C19 (s), of the
slow ligation for M5C19 (d); inset in (A) ligation turnover of C24. (B)
Equilibrium concentrations of ATP forms were calculated using the K
d
values in Table 2.
Table 2. K
d
values used to calculate equilibrium concentrations of ATP
forms. Concentration of the nicked dsDNA, 1 l
M
(0.12 m

M
DNA
phosphorus); ATP, 1 m
M
; T4 DNA ligase, 0.1 l
M
in the case of
M5C19 (8 n
M
in case of C24).
Reaction K
d
(
M
) Reference
H
+
+ ATP
4–
« HATP
3–
2.7 · 10
)7
[38]
Mg
2+
+ HATP
3–
« MgHATP
1–

6.6 · 10
)3
Mg
2+
+ ATP
4–
« MgATP
2–
8.9 · 10
)6
Mg
2+
+ MgATP
2–
« Mg
2
ATP° 1.7 · 10
)2
0.6Mg + dsDNA-PO
4
« Mg
0.6
dsDNA-PO
4
5 · 10
)6
[39,40]
Table 1. Kinetic and thermodynamic parameters of T4 DNA ligase catalysis (at 20 °C, 5 m
M
Mg

2+
and pH = 7.6, [ATP] = 1m
M
and [nds-
DNA] = 1 l
M
). Thermodynamic parameters were calculated by fitting Eqn (6) to the experimental data points, taking the transmission coefficient
j ¼ 1.
Reaction Substrate Rate, t
)1
E
A
(kcalÆmol
)1
)
DS
(calÆdeg
)1
Æmol
)1
)
Self-adenylation Mg
2
ATP 12.5 ± 0.2 s
)1
16.7 ± 0.3 1 ± 0.6
Transadenylation M5C19 1 ± 0.1 min
)1
0.9 ± 0.1 ) 65.9 ± 0.2
End-joining C6

a
83.7 ± 1.9 min
)1
16.3 ± 1.3 ) 4.3 ± 2.5
C24 34.3 ± 0.3 min
)1
16.4 ± 0.4 ) 6.1 ± 1.3
M5C19
b
4.7 ± 0.2 · 10
)3
min
)1
18.8 ± 0.7 ) 15.7 ± 2.8
a
Mean values.
b
Value relates to the slow ÔnonprocessiveÕ ligation.
Scheme 1. Kinetic scheme of self-adenylation
of T4 DNA ligase.
4320 A. V. Cherepanov and S. de Vries (Eur. J. Biochem. 270) Ó FEBS 2003
T4 DNA ligase binds ATPÁMg noncovalently, but not
ATPÁMg
2
. Subsequently, EÁATPÁMg binds a second
Mg
2+
ion, forming a catalytic intermediate, EÁATPÁMg
2
.

ATPÁMg
2
is the true substrate in the adenylation reaction,
while the monomagnesium pyrophosphate form, MgÁP
2
O
7
isthetruesubstrateforthereversereaction,thesynthesisof
ATP [24].
In this work, as part of an initial thermodynamic
characterization of T4 DNA ligase, the activation energy
of the adenylation reaction (cleavage of ATP) was deter-
mined in the absence of dsDNA using the stopped-flow
instrument. To minimize the contribution of the reverse
reaction to the observed reaction rate, we excluded pyro-
phosphate from the reaction, mixing the EÁATP complex
with Mg
2+
. In the absence of excess pyrophosphate,
k
app
À2
contributes negligibly to k
obs
2
, and k
obs
2
% k
app

2
, and
2
E
A
(obs) %
2
E
A
(app) (for the k abbreviations see Scheme 1;
for the values of k¢ see [24]).
In our experiments we used Tris/HCl buffer, which is
known for its pronounced temperature/pH-dependence
()0.026 pH units per °C). To determine the activation energy
of self-adenylation, the reaction temperature was varied
between 10 and 30 °C, and pH therefore drifted by
% 0.5 unit. We avoided using different buffering systems
because the kinetic parameters depend on the choice
of buffer. For example, the use of Tris/maleimide instead of
Tris decreases the adenylation rate % 1.5-fold (not shown).
pH-dependence
It is known that the adenylation of T4 DNA ligase strongly
depends on pH [18], because of the protonation of the
catalytic Lys159. To take into account the temperature-
induced pH drift of Tris/HCl buffer, we determined k
obs
2
at a
fixed temperature (20 °C) and different pH (6–9.5). In this
set of experiments, we used Tris/HCl buffer in part out of its

useful pH range, and special care was taken to measure pH
directly in order to account for the weakly buffering
components such as ATP, ligase and Mg
2+
(see Experi-
mental procedures).
The stopped-flow traces recorded at different pH and
fitted values of k
obs
2
are shown in Fig. 6. Eqn (4) was fitted to
the experimental data:
k
obs
2
ðpH
x
Þ%k
app
2
pH
x
ðÞ¼k
app
2
K
a
K
a
þ 10

ÀpH
ð4Þ
(pH
x
) is the pH-dependent rate constant of the forma-
tion of the enzyme–AMP adduct, k
app
2
is the pH-inde-
pendent rate constant, and K
a
is the protonation
constant of the 6-ammonium group of the catalytic
lysine residue.
Interestingly enough, the determined pK
a
¼ 9.8 ± 0.3
for the protonation of the Lys159 is more than 1.2 pH units
higher than obtained in the equilibrium binding studies [18]
and 1 pH unit lower than the pK
a
for the e-amino group of
lysine in solution [28]. This discrepancy between the results
in [18] and our data could possibly be explained by the
difference in the experimental conditions: buffer system
(Tris/HCl in our case vs. Ches, Taps and Hepes for [18]);
concentration of Mg
2+
(5 m
M

vs. 1 m
M
); reaction tem-
perature (20 °Cvs.25°C).
In the next set of experiments, we determined the k
app
2
values at different temperatures (10–30 °C) and pH ¼ 7.6
at 20 °C. Corresponding kinetic traces are shown in Fig. 7.
To correct the k
app
2
values obtained for the temperature-
induced pH drift, we employed the results of pH-depend-
ence studies shown in Fig. 6. The relation used for this
purpose was as follows:
Fig. 6. pH-dependence of the self-adenylation of T4 DNA ligase. Left:
kinetic traces obtained at different pH (values are shown next to each
trace). Right: the corresponding values of the observed rate constant
k
obs
2
. The solid trace was computed by weighted fitting using Eqn (4),
yielding values for the apparent pK
a
for the catalytic lysine residue and
pH-independent k
app
2
showninthegraph.Thereactionwasstartedby

rapid mixing of Mg
2+
solution with EÁATP complex pre-equilibrated
at the desired pH, resulting in the following final concentrations
of components: 2.7 ± 0.1 l
M
ligase, 67.35 ± 0.07 l
M
ATP,
5±0.05m
M
Mg
2+
and pH values of 6.9, 7.05, 7.32, 7.48, 7.76, 7.99,
8.28, 8.6, 8.79 and 9.2.
Fig. 7. Kinetics of the self-adenylation of T4 DNA ligase. Fluorescence
emission traces were recorded at different temperatures (values are
shown in the graph). The reaction was started by rapid mixing of
Mg
2+
solution with EÁATP complex pre-equilibrated in buffer A for
5 min at temperatures of 11.3, 13, 14.8, 16.5, 18.4, 20.2, 22.1, 24, 26,
and 27.6 °C, resulting in final concentrations of components:
2.6 ± 0.1 l
M
ligase, 71.42 ± 0.06 l
M
ATP, 5 ± 0.05 m
M
Mg

2+
.
Ó FEBS 2003 Kinetics of nick sealing by T4 DNA ligase (Eur. J. Biochem. 270) 4321
k
app
2
ðT
x
Þ
k
app
2
20

CðÞ
¼
K
a
þ 10
À7:6
K
a
þ 10
À7:6þ0:026ðT
x
À20Þ
ð5Þ
k
app
2

(T
x
) is the apparent rate constant at certain tem-
perature T
x
, k
app
2
(20 °C) is the value at 20 °C, 7.6 is the
pH of buffer A at 20 °C, and 0.026 is the pH drift of
Tris/HCl buffer per °C.
The Arrhenius plot of k
app
2
, corrected for the temperature-
induced pH changes is shown in Fig. 8. The following
equation was used to fit the experimental data points:
ln k
app
2
ÀÁ
¼ ln j
ekT
h

À
DS
#
2
R

À
2
E
A
RT
ð6Þ
The plot is essentially linear with the slope of 16.7 ±
0.3 kcalÆmol
)1
(Table 1). Combining this value with the
reaction constants determined in [24], the Mg
2+
-depend-
ent and pH-independent rate constant of adenylyl
transfer at 37 °C can be estimated as 10
3
)10
4
s
)1
. This
indicates that at physiological pH, temperature and
[Mg
2+
], T4 DNA ligase binds ATP under strongly
suboptimal conditions, resulting in two orders of mag-
nitude lower reaction rates.
Partial rates and activation energies of T4 DNA ligase
catalysis
Joining of nicked dsDNA by T4 DNA ligase involves three

catalytic steps: formation of the enzyme–adenylate, forma-
tion of the ndsDNA–adenylate, and sealing of the nick.
These processes were studied at several temperatures, to
estimate the activation energies of the individual reaction
steps under the assumption that binding of the substrate(s)
and dissociation of the product(s) do not limit the rate of
catalysis. Formation of the E–AMP intermediate was
studied using the stopped-flow technique. Synthesis of
AMP–ndsDNA and sealing of the nick were monitored
using [
32
P]ATP and/or Cy5-labeled DNA. All three proces-
ses yield essentially linear Arrhenius plots in the temperature
range +4 to +30 °C (Fig. 8). Nonlinearity for the joining
of the mismatching oligonucleotide M5C19 at high tem-
peratures may be the result of the melting of the duplex
AMP–M5C19ÁBÁC.
Formation of the E–AMP intermediate is the fastest
measured process at all temperatures studied with a rate
constant more than 10-fold higher than sealing of the
complementary ndsDNA. Both adenylation of T4 DNA
ligase and sealing of the complementary nick have reason-
ably high activation energies, which are identical within the
experimental error (Table 1). In terms of the transition-state
theory, the marked differences between the observed rates
of these reactions (i.e. adenylation of the ligase and nick
sealing) arise because of differences in the activation
entropies (Table 1). The apparent activation energy for
transadenylation of ndsDNA is % 15 kcalÆmol
)1

lower than
for adenylation of the ligase and/or sealing of the nick
(Table 1; determined for the mismatching nick
M5C19ÁBÁC). It implies that the transfer of AMP to the
terminal DNA phosphate is a thermodynamically sponta-
neous reaction, and that the T4 DNA ligase–AMP complex
is a high-energy intermediate, as suggested in [18].
Kinetics of T4 DNA ligase catalysis
In a simple description, the nick-joining activity of T4 DNA
ligase is a three-step enzymatic reaction which involves
ndsDNA, ATP and an inorganic cofactor Mg
2+
.The
reaction proceeds according to a Ping-Pong mechanism via
formation of two intermediate products: E–AMP and
AMP–ndsDNA [1,10].
In this work we observed the following phenomena: (a)
biphasic kinetics of the nick-sealing, especially pronounced
at low [ATP] (Figs 1, 2 and 4); (b) increase in the amplitude
of the burst-ligation phase at low [ATP] (Fig. 4); (c)
different [Mg
2+
]-dependence of each kinetic phase
(Fig. 5); (d) decrease in the end-joining rate at high and/or
low [ATP] (Fig. 3). To take these observations into account,
we produced Scheme 2.
Biphasic kinetics. According to Scheme 2, the initial burst-
ligation phase results from the ÔprocessiveÕ ligation (route
1
-…-6–7). In parallel to the ÔprocessiveÕ ligation, a fraction

of ligase molecules enters a nonproductive adenylation cycle
(route 1
-…-6–1), aborting the catalysis between the steps of
transadenylation and nick-sealing. Abortive adenylation
leads to the build-up of the AMP–ndsDNA pool and to the
Fig. 8. Arrhenius plot of the individual steps of T4 DNA ligase catalysis:
self-adenylation of the enzyme, transadenylation of the nick, and the end-
joining. Trace 1, self-adenylation of the ligase. The values of the
observed rate constant k
obs
2
were corrected for the temperature-induced
pH drift of the Tris/HCl buffer using Eqn (5). Traces 2, joining of C6 to
72mer/24mer BÁC. C6 to (BÁC)ratiosare1:1;3:1;10:1;30:1,and
100 : 1. Trace 3, joining of C24 to 72mer/24mer BÁC. C24 to (BÁC)
ratio is 30 : 1. Trace 4, adenylation of M5C19 in complex with 72mer/
24mer BÁC. M5C19 to (BÁC) ratio is 2 : 1. Trace 5, joining of M5C19
to 72mer/24mer BÁC. M5C19 to (BÁ C) ratio is 2 : 1. The dotted traces
were obtained by fitting Eqn (6) to the experimental data points. In the
case of trace 3 (joining of C24), and trace 5 (joining of M5C19), several
data points obtained at high temperatures were omitted to account for
melting of the DNA duplex.
4322 A. V. Cherepanov and S. de Vries (Eur. J. Biochem. 270) Ó FEBS 2003
slowing of the ligation rate, because the adenylated enzyme
(4) does not seal preadenylated DNA. The slow ligation
phase starts when the concentration of AMP–ndsDNA
reaches its maximum (steady-state conditions, Fig. 2,
20–100 min).
TherelativeratiooftheÔprocessiveÕ nick-sealing and the
abortive adenylation depends on the quality of the ndsDNA

substrate. In the case of a complementary nick, ligase forms
high-affinity complexes with the adenylated ndsDNAs [5],
and the kinetic contribution of the adenylation cycle to the
overall ligation is minor. EÁAMP–ndsDNA complexes are
less stable in the case of a 5¢-mismatching nick [12,14], and
the contribution of the adenylation cycle pathway is more
significant.
For example, the complex of ligase with AMP-
M5C19ÁBÁC is unstable, transadenylation of M5C19 is
rapid (% 1min
)1
), and the burst ligation is slow (0.2 min
)1
)
(Table 1). As a result, the slow ligation phase starts only
when all available ndsDNA substrate is converted: it is
either adenylated (via 5–6–1), or joined (via 5–6–7), and, at
the same time, the dominant enzyme fraction (4)isAMP
bound (Fig. 2, 20–100 min).
The adenylation of ligase is reversible: there is always a
fraction of the free enzyme (1) in the reaction mixture
formed via the routes 4–3–2–1, or 4–11–1). According to
Scheme 2, the slow ligation of M5C19 is performed by this
enzyme fraction via the route 1–6–7.
Increase in amplitude of the burst phase at low
[ATP]. At high [ATP] (1–5 m
M
), free ligase (1) binds
ATP (1–2) faster than it binds DNA (1–6) [(1–5) · 10
3

s
)1
for ATP vs. 10
2
s
)1
for 1 l
M
ndsDNA]. The repetitive
abortive cycling (1–2
-…-6–1) leads to accumulation of the
AMP–ndsDNA intermediate, and its removal via the
ÔnonprocessiveÕ route 1–6–7 is kinetically insignificant.
ÔNonprocessiveÕ ligation 1–6–7 becomes significant only
during the slow ligation phase, when all available ndsDNA
substrate is adenylated, and the complexes 2–5 are no longer
productive.
At low [ATP] (< 40 l
M
), ligase binds ATP slower than
DNA (< 36 s
)1
for ATP vs. 10
2
s
)1
for 1 l
M
ndsDNA).
During burst ligation, the free ligase (1) is thus engaged in

both ÔprocessiveÕ ligation (1–2
-…-6–7)andtheÔnonproces-
siveÕ scavenging of the preadenylated nicks via the route
1–6–7. ÔNonprocessiveÕ nick sealing (1–6–7) slows down the
accumulation of AMP–ndsDNA, and the start of the slow
ligation is delayed. Delay of the slow ligation results in an
increase in the amplitude of the burst-ligation phase (for
M5C19 more than 10-fold; Fig. 4). Similar results were
recently reported [15] when a large number of the dsDNA
substrates with one or two mismatching base pairs on both
sides of the nick opposite tandem canonical bases were
tested for ligation by T4 DNA ligase. There, the highest
ligation efficiency was observed at [ATP] of 10–100 l
M
.
[Mg
2+
]-dependence. The fact that the two ligation phases
have different [Mg
2+
] optima stems, in terms of Scheme 2,
from the [Mg
2+
]-regulated redistribution between the two
enzyme forms: the adenylated ligase (4), and the free enzyme
(1). E–AMPÁMg (4) is engaged in the ÔprocessiveÕ ligation
(4
-…-7). An increase in [Mg
2+
] stimulates self-adenylation

3 fi 4, and inhibits the reverse reaction 4 fi 3 [24],
causing the increase in 4, and, accordingly, the increase in
the burst-ligation rate. On the other hand, slow
ÔnonprocessiveÕ ligation is performed by the free ligase (1)
(route 1–6–7). The reverse reaction 4 fi 3 is the most
efficient at % 1–3 m
M
[Mg
2+
] [24], causing the increase in 1.
As a result, the rate of the slow phase reaches its maximum
at this [Mg
2+
]. When mismatching nicks containing tandem
canonical bases at the site of ligation are sealed, the same
narrow optimum of [Mg
2+
]between1and3m
M
has been
reported [15].
Decrease in end-joining rate at high and/or low [ATP].
According to Scheme 2, the inhibition of ligation at high
[ATP] occurs because T4 DNA ligase binds the nucleotide
at the dsDNA-binding site [20] with K
d
between 0.1 and
0.25 m
M
[21] (routes 2–9, 3–10, and 4–11).Thedecreasein

the rate of ligation at low [ATP] occurs because ndsDNA
forms a complex with the ligase (route 1–8). From the
modeling studies [26,27] one may conclude that, if the ligase
in complex (8) binds ATP at all, it will do so at a reduced
rate. The structural considerations are that ndsDNA in
complex with T4 DNA ligase prevents access of ATP to the
nucleotide-binding pocket of the enzyme, preventing ATP
from either leaving or binding to the active site. This is
reflected in Scheme 2: ligase in complex with (n)dsDNA
does not bind ATP at all.
Scheme 2. Nick-joining by T4 DNA ligase.
The rate constants k
1
–k
3
correspond to the
three steps of covalent catalysis.
Ó FEBS 2003 Kinetics of nick sealing by T4 DNA ligase (Eur. J. Biochem. 270) 4323
Scheme 2 only quantitatively addresses the AMP-
dependent reversal or inhibition of ligation. Another
important assumption is that the ssDNA fragments to be
joined do not dissociate from the opposite DNA strand or
the EÁdsDNA complex. The latter is certainly not the case
when short (4–12-mer) oligonucleotides are joined near their
T
m
. We further assumed that ndsDNA is present in the
Mg
2+
-coordinated form, neglecting the exchange between

the dsDNA-bound Mg
2+
and K
+
or Na
+
at elevated
concentrations of the latter ions (> 100 m
M
). High con-
centrations of K
+
and Na
+
inhibit DNA ligases [10,12,
29–31], perhaps, among other reasons, because ndsDNA
changes to the K
+
-(Na
+
)-coordinated form.
In summary, this kinetic scheme of T4 DNA ligase
catalysis includes a two-metal-ion mechanism of ligase
adenylation, binding of the second nucleotide molecule at
the DNA-binding site, and synthesis of dinucleoside tetra-
phosphates, and treats the reaction in terms of ÔprocessiveÕ
burst ligation and ÔnonprocessiveÕ nick sealing.
Physiological relevance of joining of mismatching
nicks by T4 DNA ligase
In vitro pre-steady-state kinetic studies performed in this

work suggest that T4 DNA ligase could rapidly adenylate
mismatching nicks in the infected cells, either at the late
stages of degradation of the cellular DNA or during the
subsequent replication of the coliphage T4 DNA: the ligase
gene is transcribed at steady levels throughout the eclipse,
reaching its maximum 3 min after infection [32]. The sealing
of these pre-adenylated nicks, however, would be slow
because of relatively high intracellular [ATP] (1–3 m
M
)[33].
Under these circumstances, T4 DNA ligase would act on
the mismatching nicks more like an mRNA capping
enzyme, a member of the same superfamily of nucleotidyl-
transferases [34,35]. In the cell, capping of the 5¢-end of
mRNA ensures protection from degradation by specific
exonucleases [36,37]. In the case of capping the mismatching
nick, however, similar reasoning does not seem logical. One
should consider that, in the cell at relatively high ionic
strength, both adenylation and joining of the mismatching
nicks could be suppressed to a large extent, as in the case of
in vitro ligations at % 0.2
M
NaCl [12,16,17]. Without
supportive experimental data in vivo, we will refrain from
assigning any physiological meaning to the ÔlowÕ substrate
specificity of T4 DNA ligase reported here and in earlier
contributions [14,15,23]. Instead, we demonstrate that this
ÔunconventionalÕ activity of T4 enzyme is an invaluable tool
for elucidating the general kinetic mechanism of DNA ligase
catalysis.

Conclusion
T4 DNA ligase-promoted end joining of nicked DNA is a
superimposition of two processes. During the burst phase,
the main enzyme fraction performs ligation ÔprocessivelyÕ, i.e.
by transadenylating ndsDNA and sealing the nick without
dissociation from the complex. In parallel, a fraction of the
ligase molecules dissociates after transadenylation, and
rebinds ATP. The slow ligation starts when most of the
ligase is AMP bound and the concentration of adenylated
ndsDNA reaches its maximal steady-state value. The end
joining of AMP–ndsDNA during the slow phase is per-
formed by a small fraction of the nonadenylated enzyme in a
ÔnonprocessiveÕ mode. The decrease in the rate of nick sealing
at low [ATP] occurs because dsDNA prevents binding of
ATP to the ligase. On the other hand, at low [dsDNA] and
high [ATP], ATP inhibits binding of ndsDNA and subse-
quent ligation by occupying the DNA-binding site.
Acknowledgements
We thank Professor W. R. Hagen for critically reading the
manuscript, and Dr P. P. Cherepanov for assistance with the
32
P
experiments. This work was supported by the Association of
Biotechnology Centers in the Netherlands (ABON) (Project I.2.8)
and in part by the Netherlands Research Council for Chemical
Sciences (CW) with financial aid from the Netherlands Technology
Foundation (STW) (grant 349-3565).
References
1. Lehman, I.R. (1974) DNA ligase: structure, mechanism, function.
Science 186, 790–797.

2. Raae, A.J. & Kleppe, K. (1978) T4 polynucleotide ligase catalyzed
joining on triple-stranded nucleic acids. Biochemistry 17, 2939–
2942.
3. Nilsson, S.V. & Magnusson, G. (1982) Sealing of gaps in duplex
DNAbyT4DNAligase.Nucleic Acids Res. 10, 1425–1437.
4. Bogenhagen, D.F. & Pinz, K.G. (1998) The action of DNA ligase
at abasic sites in DNA. J. Biol. Chem. 273, 7888–7893.
5. Rossi, R., Montecucco, A., Ciarrocchi, G. & Biamonti, G. (1997)
Functional characterization of the T4 DNA ligase: a new
insight into the mechanism of action. Nucleic Acids Res. 25,
2106–2113.
6. Weiss, B., Jacquemin-Sablon, A., Live, T.R., Fareed, G.C. &
Richardson, C.C. (1968) Enzymatic breakage and joining of
deoxyribonucleic acid. VI. Further purification and properties of
polynucleotide ligase from Escherichia coli infected with bacterio-
phage T4. J. Biol. Chem. 243, 4543–4555.
7. Weiss, B., Thompson, A. & Richardson, C.C. (1968) Ezymatic
breakage and joining of deoxyribonucleic acid. VII. Properties of
the enzyme-adenylate intermediate in the polynucleotide ligase
reaction. J. Biol. Chem. 243, 4556–4563.
8. Gumport, R.I. & Lehman, I.R. (1971) Structure of the DNA
ligase-adenylate intermediate: lysine (e-amino)-linked adeno-
sine monophosphoramidate. Proc. Natl Acad. Sci. USA 68,
2559–2563.
9. Harvey, C.L., Gabriel, T.F., Wilt, E.M. & Richardson, C.C.
(1971) Enzymatic breakage and joining of deoxyribonucleic
acid. IX. Synthesis and properties of the deoxyribonucleic acid
adenylate in the phage T4 ligase reaction. J. Biol. Chem. 246,
4523–4530.
10. Raae, A.J., Kleppe, R.K. & Kleppe, K. (1975) Kinetics and effect

of salts and polyamines on T4 polynucleotide ligase. Eur J. Bio-
chem. 60, 437–443.
11. Harada, K. & Orgel, L.E. (1993) Unexpected substrate specificity
of T4 DNA ligase revealed by in vitro selection. Nucleic Acids Res.
21, 2287–2291.
12. Wu, D.Y. & Wallace, R.B. (1989) Specificity of the nick-closing
activity of bacteriophage T4 DNA ligase. Gene 76, 245–254.
13. Pritchard, C.E. & Southern, E.M. (1997) Effects of base mis-
matches on joining of short oligodeoxynucleotides by DNA
ligases. Nucleic Acids Res. 25, 3403–3407.
14. Cherepanov, A., Yildirim, E. & de Vries, S. (2001) Joining of short
DNA oligonucleotides with base pair mismatches by T4 DNA
Ligase. J. Biochem. (Tokyo). 129, 61–68.
4324 A. V. Cherepanov and S. de Vries (Eur. J. Biochem. 270) Ó FEBS 2003
15. Alexander, R.C., Johnson, A.K., Thorpe, J.A., Gevedon, T. &
Testa, S.M. (2003) Canonical nucleosides can be utilized by T4
DNA ligase as universal template bases at ligation junctions.
Nucleic Acids Res. 31, 3208–3216.
16. Pyshnyi, D.V., Krivenko, A.A., Lokhov, S.G., Ivanova, E.M.,
Dymshits, G.M. & Zarytova, V.F. (1998) Interaction of short
oligonucleotide derivatives with nucleic acids. VI. Discrimination
of mismatch-containing complexes upon ligation of a short
oligonucleotide tandem on DNA template. Bioorg. Khim. 24,
32–37.
17. Pyshnyi, D.V., Krivenko, A.A., Lokhov, S.G., Ivanova, E.M.,
Dymshits, G.M. & Zarytova, V.F. (1998) Interaction of short
oligonucleotide derivatives with nucleic acids. V. Ligation of short
oligonucleotides in tandem on a complementary DNA template.
Bioorg. Khim. 24, 25–31.
18. Arabshahi, A. & Frey, P.A. (1999) Standard free energy for the

hydrolysis of adenylylated T4 DNA ligase and the apparent pKa
of lysine 159. J. Biol. Chem. 274, 8586–8588.
19. Madrid, O., Martin, D., Atencia, E.A., Sillero, A. & Gunther
Sillero, M.A. (1998) T4 DNA ligase synthesizes dinucleoside
polyphosphates. FEBS Lett. 433, 283–286.
20. Sillero, A. & Sillero, M.A. (2000) Synthesis of dinucleoside poly-
phosphates catalyzed by firefly luciferase and several ligases.
Pharmacol. Ther. 87, 91–102.
21. Cherepanov, A.V. & de Vries, S. (2001) Binding of nucleotides by
T4 DNA ligase and RNA ligase: optical absorbance and fluores-
cence studies. Biophys. J. 81, 3545–3559.
22. Cherepanov, A.V. & de Vries, S. (2002) Dynamic mechanism
of nick recognition by DNA ligase. Eur. J. Biochem. 269, 5993–
5999.
23. Cherepanov, A.V. & de Vries, S. (2002) Scanning mutagenesis
using T4 DNA ligase and short degenerate DNA oligonucleotides
containing tri-nucleotide mismatches. J. Biochem. (Tokyo) 132,
143–147.
24. Cherepanov, A.V. & de Vries, S. (2002) Kinetic mechanism of the
Mg
2+
-dependent nucleotidyl transfer catalyzed by T4 DNA and
RNA ligases. J. Biol. Chem. 277, 1695–1704.
25. Silber, R., Malathi, V.G. & Hurwitz, J. (1972) Purification and
properties of bacteriophage T4-induced RNA ligase. Proc. Natl
Acad. Sci. USA 69, 3009–3013.
26. Doherty, A.J. & Dafforn, T.R. (2000) Nick recognition by DNA
ligases. J. Mol. Biol. 296, 43–56.
27. Doherty, A.J. & Suh, S.W. (2000) Structural and mechanistic
conservation in DNA ligases. Nucleic Acids Res. 28, 4051–4058.

28. Fersht, A. (1999) Structure and mechanism in protein science. In A
Guide to Enzyme Catalysis and Protein Folding, p. 3. W.H. Free-
man, New York.
29. Hayashi, K., Nakazawa, M., Ishizaki, Y. & Obayashi, A. (1985)
Influence of monovalent cations on the activity of T4 DNA ligase
in the presence of polyethylene glycol. Nucleic Acids Res. 13,
3261–3271.
30. Tong, J., Cao, W. & Barany, F. (1999) Biochemical properties of a
high fidelity DNA ligase from Thermus species AK16D. Nucleic
Acids Res. 27, 788–794.
31. Lim, J.H., Choi, J., Han, S.J., Kim, S.H., Hwang, H.Z., Jin, D.K.,
Ahn, B.Y. & Han, Y.S. (2001) Molecular cloning and character-
ization of thermostable DNA ligase from Aquifex pyrophilus,a
hyperthermophilic bacterium. Extremophiles 5, 161–168.
32. Witmer, H., Baros, A., Forbes, J., Padnos, D., Maricondia, W. &
Weiner, M. (1976) Transcriptional control of T4 coliphage-specific
genes 30, 42, 43, RIIA, RIIB, and e. J. Gen. Virol. 31, 289–302.
33. Traut, T.W. (1994) Physiological concentrations of purines and
pyrimidines. Mol. Cell. Biochem. 140, 1–22.
34. Wang, S.P., Deng, L., Ho, C.K. & Shuman, S. (1997) Phylogeny
of mRNA capping enzymes. Proc. Natl Acad. Sci. USA 94,
9573–9578.
35. Shuman, S. & Schwer, B. (1995) RNA capping enzyme and DNA
ligase: a superfamily of covalent nucleotidyl transferases. Mol.
Microbiol. 17, 405–410.
36. Schwer, B., Mao, X. & Shuman, S. (1998) Accelerated mRNA
decay in conditional mutants of yeast mRNA capping enzyme.
Nucleic Acids Res. 26, 2050–2057.
37. Shatkin, A.J. (1976) Capping of eucaryotic mRNAs. Cell 9,
645–653.

38. Frey, C.M., Banyasz, J.L. & Stuehr, J.E. (1972) Interactions of
divalent metal ions with inorganic and nucleoside phosphates. II.
Kinetics of magnesium (II) with HP
3
O
4 À
10
, ATP, CTP,
HP
2
O
3 À
7
, ADP, and CDP. J. Am. Chem. Soc. 94, 9198–9204.
39. Shack, J. & Bynum, B.S. (1959) Determination of the interaction
of deoxyribonucleate and magnesium ions by means of a metal ion
indicator. Nature (London) 184, 635–636.
40. Cavalieri, L.F. (1951) Studies on the structure of nucleic acids. V.
On the mechanism of metal enzyme interactions. J. Am. Chem.
Soc. 74, 1242–1247.
Ó FEBS 2003 Kinetics of nick sealing by T4 DNA ligase (Eur. J. Biochem. 270) 4325