Tài liệu Báo cáo khoa học: Mg2+-modulated KMnO4 reactivity of thymines in the open transcription complex reflects variation in the negative electrostatic potential along the separated DNA strands ppt

Mg
2+
-modulated KMnO
4
reactivity of thymines in the open
transcription complex reflects variation in the negative
electrostatic potential along the separated DNA strands
Footprinting of Escherichia coli RNA polymerase complex
at the kP
R
promoter revisited
Tomasz Łozin
´
ski and Kazimierz L. Wierzchowski
Department of Biophysics, Institute of Biochemistry and Biophysics, Polish Academy of Sciences, Warszawa, Poland
Transcription initiation in prokaryotes involves specific
recognition between the )10 and )35 conserved hexa-
mers of the promoter DNA and RNA polymerase
holoenzyme followed by large and concerted conform-
ational changes in both components of the binary
complex leading to separation of the template and
nontemplate strands from )11 to % +4 bp, relative to
the transcription start point, and formation of the
Keywords
Escherichia coli RNA polymerase; open
transcription complex; permanganate
footprinting; thymine oxidation; kP
R
promoter.
Correspondence
K. L. Wierzchowski, Department of

Biophysics, Institute of Biochemistry and
Biophysics, Polish Academy of Sciences,
Pawin
´
skiego 5a, 02-106 Warszawa, Poland
Fax: +48 22 658 3646
Tel: +48 22 658 4729
E-mail:
Website:
(Received 24 January 2005, revised 29
March 2005, accepted 6 April 2005)
doi:10.1111/j.1742-4658.2005.04705.x
There is still a controversy over the mechanism of promoter DNA strand
separation upon open transcription complex (RPo) formation by Escheri-
chia coli RNA polymerase: is it a single or a stepwise process controlled
by Mg
2+
ions and temperature? To resolve this question, the kinetics
of pseudo-first-order oxidation of thymine residues by KMnO
4
in the
)11 … +2 DNA region of RPo at the kP
R
promoter was examined under
single-hit conditions as a function of temperature (13–37 °C) in the absence
or presence of 10 mm MgCl
2
. The reaction was also studied with respect
to thymidine and its nucleotides (TMP, TTP and TpT) as a function of
temperature and [MgCl

2
]. The kinetic parameters,
ox
k and
ox
E
a
, and
Mg-induced enhancement of
ox
k proved to be of the same order of magni-
tude for RPo–kP
R
and the nucleotides. Unlike the complex,
ox
E
a
for the
nucleotides was found to be Mg-independent. The isothermal increase in
ox
k with increasing [Mg
2+
] was thus interpreted in terms of a simple model
of screening of the negative charges on phosphate groups by Mg
2+
ions,
lowering the electrostatic barrier to the diffusion of MnO
4
–
anions to the

reactive double bond of thymine. Similar screening isotherms were deter-
mined for the oxidation of two groups of thymines in RPo at a consensus-
like Pa promoter, differing in the magnitude of the Mg effect. Together,
the findings show that: (a) the two DNA strands in the )11…+2 region of
RPo–kP
R
are completely separated over the whole range of temperatures
investigated (13–37 °C) in the absence of Mg
2+
(b) Mg
2+
ions induce an
increase in the rate of the oxidation reaction by screening negatively
charged phosphate and carboxylate groups; and (c) the observed thymine
reactivity and the magnitude of the Mg effect reflect variation in the
strength of the electrostatic potential along the separated DNA strands, in
agreement with the current structural model of RPo.
Abbreviations
R or RNAP, RNA polymerase; P, promoter; RPo, open transcription complex; Thd, thymidine; TpT, dithymidine (3¢-5¢)-monophosphate.
2838 FEBS Journal 272 (2005) 2838–2853 ª 2005 FEBS
open complex (RPo) capable of specific binding of
NTP substrates and synthesis of nascent RNA [1].
Kinetic–mechanistic studies on the initiation of tran-
scription by Escherichia coli RNA polymerase (R) at a
number of cognate promoters (P) – kP
R
[2–6], lac UV5
[7] and T7A1 [8] – on linear DNA templates showed
that formation of RPo is a multistep process involving
at least two kinetically significant intermediates, an ini-

tial complex (called I
1
or RP
c
) and an intermediate
one (I
2
or RP
i
):
R þ P !
k
1
k
À1
I
1
!
k
2
k
À2
I
2
!
k
3
k
À3
RPo

The I
1
« I
2
step is rate limiting, characterized by
extensive conformational changes and a high free
energy of activation, at which the strand separation is
proposed to be nucleated at the )11th bp [1]. In the
next step, I
2
« RPo, the latter process is completed by
a downstream expansion of the nucleated transcription
bubble. A measurable population of the I
2
form could
be observed for the lac UV5 and kP
R
promoters on
linear DNA templates only. Negative supercoiling of
the DNA template shifts the opening equilibrium
(K
3
¼ k
3
⁄ k
-3
) towards RPo [3,7,9]. For the kP
R
pro-
moter on a negatively supercoiled plasmid, this equilib-

rium has been found shifted completely towards RPo
over the temperature range 4–37 °C [10].
According to the current molecular model of RPo
[11,12], based on crystal structures of Thermus thermo-
philus [13] and Thermus aquaticus [14] RNA
polymerase holoenzyme (free and complexed with
forked DNA), the two separated DNA strands are
held in protein channels formed by segments of the
r
70
and b, b¢ RNA polymerase subunits. The mole-
cular mechanism of DNA strand separation is still
unknown, however. Studies with mutant polymerases
harboring deletions in these subunits seem to be very
promising with respect to this problem. They have led
to selection of mutants forming partial melting inter-
mediates [15,16], identification of RNA polymerase
(RNAP) regions involved in melting, and demonstra-
tion that isolated b¢(1–314) and r
2)3
fragments alone
are able to melt an extended )10 promoter [17,18].
Also the question of whether DNA melting by
RNAP occurs as a concerted one-step or stepwise pro-
cess controlled by temperature and Mg
2+
ions [1]
remains disputable.
In view of the absolute requirement for Mg
2+

in
the process of RNA synthesis by E. coli RNAP [19],
related primarily to the involvement of Mg
2+
in the
catalysis of internucleotide phosphodiester bond for-
mation and binding of NTP substrates [20–23], a pos-
sible role of these ions in RPo formation has been
probed by kinetic [4] and footprinting experiments
[10,24–26]. A large difference between the observed
(% 0.4) and expected ( % 4; from that of % 7 found for
Na
+
[2]) stoichiometry of Mg
2+
ion uptake in the kin-
etically significant steps of RPo–kP
R
dissociation [4],
led the authors to postulate a ‘fourth step’ hypothesis
stating that in the absence of Mg
2+
an intermediate
open complex, RPo1, is formed which, upon specific
binding of three Mg
2+
ions, transforms to its tran-
scription-competent form, RPo2. Strong enhancement
by Mg
2+

of susceptibility to KMnO
4
oxidation of
pyrimidine bases at the )12, +1 and +2 positions in
kP
R
promoter DNA, found in subsequent footprinting
experiments [10], led the authors to postulate
Mg-induced expansion of the transcription bubble
from its center to both ends, accompanying the
RPo1 fi RPo2 transition. Further studies of the Mg
effect on the kP
R
DNA backbone scission by hydroxy
radicals [24], generated in the Fenton reaction between
Fe(EDTA)
2–
and H
2
O
2
, indicated that deoxyribose resi-
dues at positions close to the transcription start point
react with OH in RPo2 but are relatively protected in
RPo1, suggesting rather a downstream expansion of
the transcription bubble upon binding of Mg
2+
. It has
been proposed to result from greater steric accessibility
and ⁄ or local reduction in negative charge density asso-

ciated with DNA phosphates and carboxylates in the
catalytic pocket of E. coli RNAP holoenzyme in
RPo2. Results of a similar study on the RPo–T7A1
complex [26], based on oxidation of thymines by
KMnO
4
and OsO
4
and of DNA backbone scission by
hydroxy radicals as a function of temperature, indi-
cated that the transcription bubble consists of a
Mg-independent part and a Mg-dependent one close
to the catalytic site, both having individual transition
temperatures indicative of a stepwise expansion of the
melted DNA region. The appearance of discrete melt-
ing intermediates in the complex formed by Bacillus
subtilis RNAP at the flagellin promoter [27] was also
claimed on the basis of temperature-dependent changes
in the permanganate footprint pattern. In all these
studies, very high multiple-hit doses of the permangan-
ate were used, and the pyrimidine oxidation was
assumed to be temperature-independent. To assess
more reliably multiple-hit footprinting data, Tsodikov
et al. [28] developed a quantitative method of analysis,
in which chemical probing performed as a function of
either concentration of the oxidant or time exposure
allows evaluation of reactivity rate constants for indi-
vidual bases. Application of this method to perman-
ganate oxidation of RPo–kP
R

showed that of the three
bases (T)4, T)3 and T+2) probed at 37 °C and 0 °C,
only the reactivity of the last one was temperature-
T. Łozin
´
ski and K. L. Wierzchowski Open complex at kP
R
promoter
FEBS Journal 272 (2005) 2838–2853 ª 2005 FEBS 2839
dependent [28]. Our attempts to apply this method to
RPo at a synthetic Pa promoter failed, however,
because of the occurrence of highly competitive oxida-
tion reactions within the RNAP component [29], which
was ignored in the method. We have shown that under
multiple-hit KMnO
4
doses commonly used in the ear-
lier footprinting experiments, RPo becomes completely
inactivated, through severe damage by multiple oxida-
tive lesions accumulating in both the RNAP and the
melted DNA region, and partially dissociated. Perman-
ganate footprinting of RPo–Pa as a function of single-
hit oxidant dose [30] showed that, in this complex,
Mg
2+
ions do not induce any expansion of the melted
DNA region, but merely increase the reactivity of all
thymines in a position-dependent manner, in particular
those located close to the active center of the com-
plex, as observed previously at other promoters

[10,24,26,27]. In a parallel study of the rate of RPo–Pa
complex dissociation as a function of [Mg
2+
], we have
shown [31] that, in the 20–37 °C temperature range,
four Mg
2+
ions are involved in the equilibrium K
3
¼
k
3
⁄ k
-3
(equivalent to seven Na
+
ions found in the case
of RPo–kP
R
[2]), as could be expected for a fully
melted transcription bubble. Our current studies on
the dependence on [Mg
2+
] of the rate of RPo–kP
R
dis-
sociation show that its course is biphasic, which may
indicate involvement of ionic exchange coupled to re-
formation of salt bridges on the protein surface upon
dissociation of wrapped DNA from the complex [32].

In view of (a) the presented critical assessment of
the experimental approaches used in previous foot-
printing studies, in particular the clearly unrealistic
assumption that the underlying chemical reaction of
thymine oxidation is temperature-independent, and (b)
the possibility that the mechanism of transcription
bubble formation may depend on the promoter
sequence [1], it seemed necessary to reinvestigate per-
manganate oxidation of the most studied RPo–kP
R
complex as a function of the oxidant single-hit dose,
temperature and Mg
2+
concentration. Here we report
the results of these experiments. They clearly show that
the pattern of oxidation of thymines in the bubble
region of RPo at kP
R
at 37 °C is generally similar to
that determined previously by us for RPo at the Pa
promoter on the same template under similar condi-
tions [30], only the reactivity of thymines in the RPo–
kP
R
complex appeared to be significantly higher than
that of analogously located bases in RPo–Pa. More-
over, oxidation of thymines was shown to exhibit
temperature-dependence similar to that found for
thymidine and its nucleotides (see below).
We believe that the effect of Mg

2+
on thymine oxi-
dation in RPo–Pa [30] and in dsDNA [33] was due
mainly to the screening of negative charges of DNA
phosphates (and carboxylic groups in RPo) near the
thymine residues. The reduction of local negative
charge by Mg
2+
ions bound in the catalytic pocket of
RPo–kP
R
was also considered by the group of Record
[24] as a possible source of the increased backbone and
base reactivity at the start site. Additional arguments
in support of this notion were obtained in this work
from experiments on KMnO
4
oxidation of thymine
residues in thymidine (Thd), TMP, TTP and dithymi-
dine (3¢-5¢)-monophosphate (TpT) as a function of
temperature and [Mg
2+
].
In connection with the recent model of RPo struc-
ture [11,12], we show finally that the observed reacti-
vity of thymine and its modulation by Mg
2+
reflects
variation in the effective electrostatic potential along
the separated DNA strands determined by charged

DNA phosphates and protein groups at the walls of
RNAP channels surrounding the DNA.
Results
Oxidation of the RPo–kP
R
complex
Thymine residues in the promoter bubble DNA
region of RPo formed by E. coli RNAP at the kP
R
promoter (see Fig. 1 for sequence), contained in the
pDS3 plasmid, were oxidized by KMnO
4
in the
absence and presence of 10 mm MgCl
2
at three selec-
ted temperatures (13 °C, 25 °C and 37 °C) as a func-
tion of the oxidant dose (x ¼ [KMnO
4
] · t, m · s) in
the range 0.004–0.04 m · s, which is known to ensure
single-hit oxidation of thymines in the melted DNA
region and preservation of the original structure of
the complex almost intact [29]. At 10 mm MgCl
2
a
high occupancy of the catalytic site can be expected
on the basis of micromolar value of the apparent
equilibrium binding constant for Mg
2+

to E. coli
RNAP, which can be estimated from the protective
effect of Mg
2+
on Fe
2+
-induced cleavage of the pro-
tein fragments forming this site [22]. PAGE-resolved
DNA products of the Klenow extension reaction,
Fig. 1. Sequences of the kP
R
and Pa promoters studied. Melted
regions in RPo are shown in bold, )10 and )35 recognition hexa-
mers are underlined, and an arrow marks the transcription start
point.
Open complex at kP
R
promoter T. Łozin
´
ski and K. L. Wierzchowski
2840 FEBS Journal 272 (2005) 2838–2853 ª 2005 FEBS
corresponding to oxidized thymine, are exemplified in
Fig. 2. Inspection of the gels shows that DNA bands
corresponding to all thymines in the )11 … +2 pro-
moter region came up clearly in the footprints
obtained even at 13 °C and the lowest oxidant dose
applied and in the absence of MgCl
2
. In the foot-
prints from reactions carried out in the presence of

10 mm MgCl
2
, all these bands are merely more
intense. Note that, neither for the four cytosines pre-
sent in the bubble region (at positions )1, )2, )5 and
)6) nor for C)12, found oxidized under multiple-hit
conditions [10], can a DNA band corresponding to
oxidized base be traced even at the highest oxidant
dose applied and 37 °C.
The extent of thymine oxidation in the template and
nontemplate DNA strands was evaluated (as described
in Experimental procedures) by quantification of the
corresponding
32
P-end-labeled primer-extended DNA
fragments in the footprints. The average fractions of
oxidized thymine thus obtained,
ox
f
i
(x), rose mono-
exponentially as a function of the applied oxidant dose
x, as expected for a pseudo-first-order reaction. This is
exemplified in Fig. 3 for reactions performed at 37 °C.
The corresponding rate constants of the reaction,
ox
k
i
,
A

B
CD
Fig. 2. Selected KMnO
4
footprints of the melted DNA region of RPo at the kP
R
promoter. (A) and (B) Autoradiograms of 6% polyacrylamide
sequencing gels (at the right side the whole, and at the left side enlarged fragments corresponding to the melted DNA region) showing
resolved
32
P-end-labeled ssDNA products of the Klenow primer extension reaction carried out on the nontemplate (A) and template (B) DNA
strands; doses of KMnO
4
(in M · s) applied at 37 °C are indicated above lanes 2–11. Minus and plus signs indicate the absence and pres-
ence of 10 m
M MgCl
2
in footprinting reactions. Lane 1, footprint of dsDNA without RNA polymerase. (C) and (D) Fragments of autoradio-
grams of 6% polyacrylamide sequencing gels showing resolved
32
P-end-labeled ssDNA products of the Klenow primer extension reaction
carried out on the nontemplate (C) and template (D) DNA strands of RPo oxidized at different temperatures and KMnO
4
doses (indicated
above the lanes); along the leftmost lanes positions of DNA bands corresponding to thymines in the melted region of the kP
R
promoter are
indicated. Some bands ascribed to particular oxidized thymines are doubled: the stronger component corresponds to the DNA extension
reaction products ending at thymine diglycol, and the weaker one, to fragments shorter by one base formed when the extension reaction
encountered oxidized thymine hydrolyzed to the ureido form.

T. Łozin
´
ski and K. L. Wierzchowski Open complex at kP
R
promoter
FEBS Journal 272 (2005) 2838–2853 ª 2005 FEBS 2841
were thus obtained by nonlinear weighted fit of Eqn
(1) to the experimental data:
ox
f
i
¼ 1 À expðÀ
ox
k
i
 xÞð1Þ
They are collected in Table 1. The pseudo-first-order
character of the oxidation reaction testifies that all thy-
mines in RPo under the conditions studied are solvent
accessible and that the flux of MnO
4
–
anions to the
reaction sites within protein channels is high enough
to sustain this type of kinetics for the bimolecular
reaction.
At 37 °C and 10 mm Mg
2+
, that is under conditions
in which a transcription competent RPo–kP

R
complex
has been shown to be the dominant species with fully
separated DNA strands [10,24], the most reactive
thymines were T)11 and T)8 of the template (t)
strand and T)3 and T)4 of the nontemplate (nt)
strand, whereas T+1 (t) and T+2 (nt), located close
to the catalytic center, were much less reactive. The
least reactive, however, proved to be T)7 and T)10 of
the nt strand, which are known to be involved in
specific interactions with region 2.3 of r
70
[1] and T)9
Fig. 3. Kinetics of KMnO
4
oxidation at 37 °C
of thymines in the melted region of RPo at
the kP
R
promoter. Data points (mean val-
ues: n ¼ 3, calculated standard errors in the
range of 10–15%) corresponding to DNA
fractions of oxidized thymines (
ox
f
i
) and
unoxidized DNA (f
uDNA
) in the template (left

column), and nontemplate (right column)
strands, in the absence (j) and presence
(d)of10m
M MgCl
2
were obtained by quan-
tification of the footprints (exemplified in
Fig. 2) as described in Experimental Proce-
dures; solid lines represent fitted functions
(Eqn 1,
ox
k
i
values in Table 1).
Open complex at kP
R
promoter T. Łozin
´
ski and K. L. Wierzchowski
2842 FEBS Journal 272 (2005) 2838–2853 ª 2005 FEBS
of the t strand. The patterns of thymine reactivity in
the presence and absence of Mg
2+
were generally
similar. In the absence of Mg
2+
, the reactivity of all
the thymines was merely lower. A comparison of the
corresponding
ox

k
i
and
ox
k
i,Mg
values shows that
thymine reactivity in the presence of Mg
2+
becomes
enhanced by a position-dependent factor,
ox
k
i
,
Mg
⁄
ox
k
i
,
with the largest values of 4.4 for T+1 and 3.5 for
T+2, and much smaller, in the range 1.6–2.3, for the
most reactive groups including T)3, T)4, T)8 and
T)11. For the least reactive, T)9, the Mg effect was
similar to that of the last group, whereas for T)7 and
T)10 its value of 1.2 was distinctly smaller. The
pattern of relative thymine reactivity did not change
significantly on lowering the temperature from 37 °C
to 25 °C and 13 °C; only the rate constants of oxida-

tion became progressively smaller, as would be expec-
ted for a chemical reaction, and the Mg effect became
somewhat larger, for T+1 and T+2 in particular.
The Arrhenius plots of ln(
ox
k
i,Mg
) and ln(
ox
k
i
) vs.
1 ⁄ T (K) were linear (Fig. 4, correlation coefficient
)0.98 or better) as if
ox
k
i
reflected mainly the tempera-
ture dependence of the oxidation reaction. The ener-
gies of activation,
ox
E
a
, calculated from these plots
(Table 2) for reactions carried out in the absence of
Mg
2+
were 6–8 kcalÆmol
)1
(25.1–33.5 kJÆmol

)1
) for
T)3, T)4, T)7, T)10 and T)11, distinctly higher at
% 11 kcalÆmol
)1
(% 46 kJÆ mol
)1
) for T+2 and T+1,
and much lower at 3.7 kcalÆmol
)1
(15.5 kJÆmol
)1
) for
T)8. The higher values of
ox
E
a
for the oxidation of
T+1 and T+2, which are located close to the active
center of RPo, correlate with the lower reactivity of
these bases. For the reactions carried out in the pre-
sence of Mg
2+
, the corresponding values of
ox
E
a,Mg
proved to be generally smaller for all thymines. The
largest decrease, by a factor of % 2, was found for
T+2, and T)7 and T)10. The reactivity of T)9,

which apparently did not change with temperature in
the absence of Mg
2+
, varied in the same way as T)8
in the presence of Mg
2+
.
Table 1. Pseudo-first-order rate constants (
ox
k
i
) for thymine oxida-
tion by KMnO
4
in the melted DNA region of RPo at the kP
R
promo-
ter. The mean ± SEM values of
ox
k
i
were determined from
nonlinear weighted least squares fit of Eqn (1) to the
ox
f
i
(x) data
(Fig. 3) obtained as described in Experimental procedures. nt, Non-
template DNA strands; t, template DNA strand.
Thymine

Temperature
(°C)
ox
k
i
(M
)1
Æs
)1
)
[Mg
2+
] ¼ 0
ox
k
i,Mg
(M
)1
Æs
)1
)
[Mg
2+
] ¼ 10 mM
ox
k
i,Mg
⁄
ox
k

i
T+2 (nt) 37 3.0 ± 0.1 10.5 ± 0.2 3.5 ± 0.2
T-3 (nt) 7.0 ± 0.2 15.9 ± 0.5 2.3 ± 0.1
T-4 (nt) 6.1 ± 0.2 10.4 ± 0.4 1.7 ± 0.1
T-7 &
T-10 (nt)
0.8 ± 0.1 1.0 ± 0.1 1.25 ± 0.1
T-11 (t) 14.8 ± 0.3 23.3 ± 0.5 1.6 ± 0.05
T-9 (t) 0.9 ± 0.1 1.4 ± 0.1 1.6 ± 0.2
T-8 (t) 9.5 ± 0.2 16.8 ± 0.4 1.8 ± 0.05
T+1 (t) 1.2 ± 0.1 5.2 ± 0.1 4.3 ± 0.4
T+2 (nt) 25 1.5 ± 0.1 7.8 ± 0.1 5.2 ± 0.35
T-3 (nt) 4.0 ± 0.1 10.8 ± 0.2 2.7 ± 0.1
T-4 (nt) 3.4 ± 0.1 6.7 ± 0.2 2.0 ± 0.1
T-7 &
T-10 (nt)
0.5 ± 0.1 0.9 ± 0.1 1.8 ± 0.4
T-11 (t) 10.2 ± 0.2 19.6 ± 0.3 1.9 ± 0.05
T-9 (t) 0.9 ± 0.1 1.1 ± 0.1 1.2 ± 0.4
T-8 (t) 7.6 ± 0.1 15.5 ± 0.3 2.0 ± 0.05
T+1 (t) 0.5 ± 0.1 2.6 ± 0.1 5.2 ± 1.1
T+2 (nt) 13 0.7 ± 0.1 5.5 ± 0.1 7.9 ± 1.1
T-3 (nt) 2.7 ± 0.1 7.3 ± 0.1 2.7 ± 0.1
T-4 (nt) 2.2 ± 0.1 4.8 ± 0.1 2.2 ± 0.15
T-7 &
T-10 (nt)
0.3 ± 0.1 0.6 ± 0.1 2.0 ± 0.75
T-11 (t) 6.4 ± 0.1 13.9 ± 0.2 2.2 ± 0.05
T-9 (t) 0.9 ± 0.1 0.9 ± 0.2 1.0 ± 0.25
T-8 (t) 5.8 ± 0.1 11.7 ± 0.2 2.0 ± 0.05

T+1 (t) 0.4 ± 0.1 2.6 ± 0.2 6.5 ± 1.7
Fig. 4. Effect of temperature and Mg
2+
on the rate constants of thymine oxidation by KMnO
4
in RPo at the kP
R
promoter, in Thd and TMP.
Arrhenius plots of
ox
k
i
data (Tables 1 and 3) in the absence (j) and presence (d)of10mM MgCl
2
; solid lines represent fitted functions, cor-
responding activation energies of the reaction in Table 2.
T. Łozin
´
ski and K. L. Wierzchowski Open complex at kP
R
promoter
FEBS Journal 272 (2005) 2838–2853 ª 2005 FEBS 2843
There is no doubt that temperature also influences
the structural dynamics of RPo and may thus induce
some local conformational changes in the complex
affecting the accessibility of reaction centers to the
oxidant. This may apply for instance to T+1 and
T+2 located in a more structurally rigid motif of
RPo. The measured values
ox

E
a
should thus be regar-
ded as apparent. The values of
ox
E
a,Mg
are formally
smaller because the magnitude of the Mg effect
increases progressively as the temperature falls
(Table 1), which may be due to increased binding of
Mg
2+
. This point is dealt with further in the Discus-
sion.
Oxidation of thymine in thymidine and
its nucleotides
The kinetics of oxidation of thymine by KMnO
4
[34]
in thymidine, TMP, TTP and TpT was studied as a
function of [MgCl
2
] in the range 0–100 mm in the pres-
ence of 100 mm KCl (Figure 5 and Table 3). The
pseudo-first-order rate constants of the reaction in the
absence of Mg
2+
appeared to be of similar magnitude
to those determined for thymines in RPo–kP

R
under
similar salt and temperature conditions (Tables 1 and
2). For the nucleotides, they were smaller than for the
parent nucleoside (21.1 m
)1
Æs
)1
) and decreased with the
increase in negative charge on the phosphate group in
the order TpT (14.4 m
)1
Æs
)1
), TMP (8.5 m
)1
Æs
)1
) and
TTP (6.4 m
)1
Æs
)1
).
In agreement with the hypothesis referred to above,
the rate of Thd oxidation was independent of the
presence of Mg
2+
, whereas those of TMP, TpT and
TTP exhibited a dependence on [MgCl

2
] mimicking a
binding isotherm (Fig. 5), which was particularly steep
for TTP.
It is known that Thd in aqueous solution adopts
the anti conformation about the N(1)–C(1¢) glycosidic
Table 2. Energies of activation of the KMnO
4
-oxidation reaction
(
ox
E
a
) of thymines in the melted DNA region in RPo at the kP
R
pro-
moter, and in Thd and TMP. The mean ± SEM values of
ox
E
a
were
determined from linear weighted least-squares fit of the Arrhenius
equation to the
ox
k
i
data from Table 1 and Table 3.
Substrate
ox
E

a
(kcalÆmol
)1
)
[Mg
2+
] ¼ 0
ox
E
a,Mg
(kcalÆmol
)1
)
[Mg
2+
] ¼ 10 mM
ox
E
a,Mg
⁄
ox
E
a
T+2 (nt) 10.7 ± 0.1 4.8 ± 0.1 0.45 ± 0.03
T-3 (nt) 7.2 ± 1.0 5.7 ± 0.1 0.79 ± 0.11
T-4 (nt) 7.7 ± 0.9 5.5 ± 0.6 0.71 ± 0.23
T-7 & T-10 (nt) 7.2 ± 0.1 3.0 ± 1.7 0.42 ± 0.24
T-11 (t) 6.2 ± 0.2 4.0 ± 0.7 0.64 ± 0.11
T-9 (t) 0 3.5 ± 0.3 –
T-8 (t) 3.7 ± 0.1 2.8 ± 0.8 0.78 ± 0.22

T+1 (t) 11.1 ± 3.3 7.2 ± 3.2 0.65 ± 0.35
Thd 8.6 ± 0.3 8.4 ± 0.3 0.98 ± 0.11
TMP 7.7 ± 0.2 8.1 ± 0.5 1.05 ± 0.07
[Mg
2+
] ¼ 50 mM
8.3 ± 0.1
Fig. 5. Effect of [MgCl
2
] on the kinetics of thymine oxidation by
KMnO
4
. In TpT (A), TMP (B), TTP (C), and in two nontemplate DNA
strand regions: T+2, T+3 (d)andT)2, T)3, T)4(j) of RPo at the
Pa promoter (D). The rate constants
ox
k of the reaction in nucleo-
tides were determined at 25 °C, and sums of the respective
ox
k
i
in
the RPo–Pa complex at 37 °C. Solid lines represent fitted functions
(Eqn 2), the values of the fitted parameters in Table 4.
Open complex at kP
R
promoter T. Łozin
´
ski and K. L. Wierzchowski
2844 FEBS Journal 272 (2005) 2838–2853 ª 2005 FEBS

bond [35] in which the C-5¢-OH hydroxy group
makes close contact with the C(6)-H group of the
thymine ring [36]. In thymidine 5¢-phosphates, the
negatively charged terminal monophosphate and tri-
phosphate groups are thus expected to be located
close to the C(5)¼C(6) double bond susceptible to
MnO
4
–
attack. In solutions close to neutrality, these
groups chelate only one Mg
2+
ion [35]. The same is
expected for the diester phosphate group in TpT. The
ox
k([Mg
2+
]) data for all these compounds were thus
fitted to a simple model assuming a single Mg
2+
-
binding site involved in the screening of negative
phosphate charges:
ox
k
Mg
¼
ox
k þ Df½Mg
2þ

ÂK
scr
=ð1 þ½Mg
2þ
ÂK
scr
Þg
ð2Þ
where K
scr
is a screening constant, expected to be pro-
portional to the corresponding thermodynamic binding
constant, K
ass
, for Mg
2+
and D ¼
ox
k
([Mg] fi 1)
–
ox
k
([Mg] ¼ 0)
is an increment by which the initial value
of
ox
k would increase at the saturating Mg
2+
concen-

tration. Values of K
scr
thus obtained at 25 °C (25.7,
32.4 and 350 m
)1
) for TpT, TMP and TTP, respect-
ively (Table 4) are presumably smaller than the
respective binding constants K
ass
, as they reflect only
replacement by Mg
2+
of K
+
ions from solvation shells
of phosphate groups leading to more effective screen-
ing of their negative charges. The ratio of K
scr
values
for TMP and TTP of % 10 is sixfold smaller than that
of the intrinsic equilibrium binding constants for UMP
and UTP [37], measured at 100 mm NaCl, which is
probably due to the multitude of conformations adop-
ted by the Mg-chelating triphosphate group [35], some
of which apparently do not contribute significantly to
the electrostatic barrier to MnO
4
–
being considered.
The corresponding values of

ox
k
([Mg] fi 1)
¼ D +
ox
k
([Mg] ¼ 0)
, that is the rate constant at saturating
Mg
2+
concentration (18.3, 15.5 and 13.5 m
)1
Æs
)1
) tend
to approach that of 21.1 m
)1
Æs
)1
measured for thymi-
dine. It is thus evident that the role played by Mg
2+
in the enhancement of the reactivity of thymine
towards MnO
4
–
in thymidine phosphates is mainly
electrostatic in nature. Therefore, the differences in
ox
k

([Mg] ¼ 0)
between thymidine and its nucleotides
reflect mostly the differences in the electrostatic barrier
to diffusion of MnO
4
–
to the reactive double bond of
the thymine moiety. The remaining differences between
the corresponding
ox
k
([Mg] fi 1)
values can be ascribed
to steric factors determined by the different sizes of the
substituents. In the case of TpT, intramolecular stack-
ing of thymine residues brings the two C(5)¼C(6)
bonds in close proximity, thereby increasing the prob-
ability of their attack by MnO
4
–
and decreasing the
effect of the negative phosphate charges on MnO
4
–
dif-
fusion by dielectric shielding.
The temperature dependence of the kinetics of oxi-
dation of Thd and TMP was also investigated, and the
Arrhenius energies of activation determined, in the
absence of Mg

2+
and the presence of selected Mg
2+
Table 3. Pseudo-first-order rate constants (
ox
k) for thymine oxidation by KMnO
4
in Thd and its nucleotides. The mean ± SEM values of
ox
k
were determined from nonlinear weighted least-squares fit of a single exponential decay function to the kinetic data obtained as described
in Experimental Procedures.
Compound
Temperature
(°C)
ox
k (M
)1
Æs
)1
)
[Mg
2+
] ¼ 0
ox
k
Mg
(M
)1
Æs

)1
)
[Mg
2+
] ¼ 10 mM
ox
k
Mg
(M
)1
Æs
)1
)
[Mg
2+
] ¼ 50 mM
Thd 25 21.1 ± 0.9 20.7 ± 0.9
13 12.0 ± 0.4 11.8 ± 0.5
1 6.0 ± 0.2 6.0 ± 0.3
TpT 25 14.4 ± 0.2 15.4 ± 0.3
TMP 37 13.5 ± 0.5 15.3 ± 0.7 20.7 ± 0.3
25 8.5 ± 0.4 10.4 ± 0.3 12.3 ± 0.2
13 5.3 ± 0.5 6.0 ± 0.5 6.8 ± 0.1
1 2.6 ± 0.1 2.9 ± 0.1 3.5 ± 0.1
TTP 25 6.4 ± 0.2 11.2 ± 0.4
Table 4. Fitted parameters of Mg
2+
-screening isotherms for
KMnO
4

oxidation of thymidine nucleotides and the two groups of
thymine residues in RPo–Pa. The parameter’s values (mean ±
SEM) were determined from nonlinear weighted least-squares fit
of Eqn (2) to the data points shown in Fig. 5.
Compound
Temperature
(°C)
K
scr
(M
)1
)
ox
k
([Mg] ¼ 0)
(M
)1
Æs
)1
)
ox
k
([Mg] fi 1)
(M
)1
Æs
)1
)
TpT 25 25.7 ± 6.6 14.4 ± 0.1 18.3 ± 0.5
TMP 32.4 ± 5.7 8.2 ± 0.1 15.5 ± 0.5

TTP 350 ± 78 6.5 ± 0.3 13.5 ± 1.1
RPo–Pa 37
ST)4, T)3, T)2 161 ± 64 9.3 ± 1.1 21.1 ± 2.7
ST+2, T+3 162 ± 12 3.0 ± 0.1 10.9 ± 0.4
T. Łozin
´
ski and K. L. Wierzchowski Open complex at kP
R
promoter
FEBS Journal 272 (2005) 2838–2853 ª 2005 FEBS 2845
concentrations (Table 2 and Fig. 4). They are of the
same order of magnitude as those determined for thy-
mine oxidation in RPo–kP
R
. The activation energies of
Thd and TMP oxidation in the absence and presence
of 10 mm MgCl
2
, and for the latter also at 50 mm
MgCl
2
, were found to be similar within the limits of
experimental error. This is an important observation,
confirming that the mechanism underlying the Mg
2+
effect on the kinetics of nucleotide oxidation does not
affect the intrinsic rate of the reaction and is due solely
to the screening of the negative charge on the phos-
phate group, thereby diminishing the electrostatic bar-
rier to diffusion of MnO

4
–
to the reactive double bond
of the thymine moiety.
Mg
2+
effect on the kinetics of oxidation of the
RPo–Pa complex
The two separated DNA strands in RPo are held in
protein channels formed by segments of the r
70
and b
RNAP subunits [11,12]. Therefore, the thymines can
be expected to experience different molecular environ-
ments depending on their location. It was thus of
interest to determine screening isotherms for variously
positioned thymine residues, analogous to those
obtained for the nucleotides in the preceding section.
For this experiment we used RPo at the consensus-like
Pa promoter [30], which at 37 °C exhibits a very sim-
ilar pattern of thymine oxidation in the ssDNA region
to that observed here for the RNAP–kP
R
complex
(Fig. 7), and, unlike RPo–kP
R
, resists relatively high
[MgCl
2
] in titration experiments [31]. RPo formed by

E. coli RNAP in a buffer solution containing 100 mm
KCl and varying [MgCl
2
] in the range 0–40 mm was
oxidized at 37 °C with a KMnO
4
dose of 0.01 or
0.02 m · s, found to be sufficiently low to secure single-
hit conditions of oxidation within the whole range of
MgCl
2
concentrations applied [29]. The corresponding
footprints (Fig. 6) were quantified, and the kinetic
parameters derived as described for the RPo–kP
R
com-
plex. Sums of the rates of oxidation, S
ox
k
i
, of the two
groups of thymines in the nontemplate promoter strand
–(a) T+3 and T+2 and (b) T)2, T)3 and T)4 found
to differ greatly in the magnitude of the Mg
2+
effect at
10 mm concentration [30] – are plotted as a function of
[MgCl
2
] in Fig. 5. The plots resemble closely the screen-

ing isotherms obtained for the nucleotides. Fitting of
Eqn (2) to these data yielded a value of K
scr
% 160 m
)1
,
similar for both groups of thymines (Table 4). This
confirms that the Mg
2+
ions involved in the screening
of negative charges associate with binding sites of sim-
ilar affinity in the two bubble regions, that is most
probably DNA phosphates. In agreement with earlier
findings [30], the two groups are characterized by very
different values of the
ox
k
([Mg] fi 1)
⁄
ox
k enhancement
factor of 3.7 and 2.3 for (a) and (b), respectively. Con-
sequently, the maximum value of S
ox
k
i([Mg] fi 1)
for the
less reactive group (a) in the absence of Mg
2+
tends to

approach that of the more reactive one (b), as observed
for TTP and TMP. The still smaller maximum reactiv-
ity of T+2 and T+3 close to the catalytic center of
RPo can be thus attributed by similar token to a larger
steric barrier to diffusion of permanganate anions to
thymines in this region. Because, even at the lowest
oxidant doses applied, the RPo was found to be com-
pletely inactivated transcriptionally [29], conformation-
al changes in the catalytic center, probably caused by
oxidation of Cys454 close to the NADFDGD motif
[29], may in part be responsible for the observed lower
steric accessibility of the two thymines to the oxidant in
both the presence and absence of Mg
2+
. This conclu-
sion also applies to RPo at kP
R
and other promoters.
Discussion
The results show that (a) all nine thymine residues of
the template and nontemplate strands in the
)11 … +2 region of RPo at a plasmid-borne kP
R
Fig. 6. Effect of [MgCl
2
] on KMnO
4
footprint of the melted DNA region in RPo at the Pa promoter. Autoradiogram of polyacrylamide sequen-
cing gel (6%) showing resolved
32

P-end-labeled ssDNA products of the Klenow primer extension reaction carried out on the nontemplate
DNA strand: at the right side whole gel, and at the left side fragment corresponding to the melted DNA region; [MgCl
2
]inmM is indicated
above the lanes; along the leftmost lane positions of DNA bands corresponding to thymines in the melted region are indicated.
Open complex at kP
R
promoter T. Łozin
´
ski and K. L. Wierzchowski
2846 FEBS Journal 272 (2005) 2838–2853 ª 2005 FEBS
promoter are susceptible to permanganate oxidation in
the temperature range 13–37 °C and (b) the corres-
ponding reaction rate constants are of the same order
of magnitude as, and exhibit similar Mg
2+
and tem-
perature dependence to, those for thymine oxidation in
free thymidine nucleotides. The simplest interpretation
of these findings is that the DNA strands in this region
are completely separated, as expected for the fully
‘open’ RPo, and the observed temperature dependence
of the oxidation rate constants can be interpreted as
being due, for the most part, to the inherent activation
energy of the reaction. Moreover, the enhancement of
thymine reactivity, observed at 10 mm Mg
2+
in both
RPo–kP
R

and RPo–Pa, was shown in RPo–Pa to be a
continuous function of Mg
2+
concentration. Together,
the results of single-hit permanganate footprinting of
RPo–kP
R
do not support the earlier interpretation of
the Mg
2+
effect in single-dose multi-hit experiments
for this complex [10], which stated that Mg
2+
induces
in a partially opened subpopulation of RPo, called
RPo1, some conformational transition leading to an
extension of the melted DNA region from the center
outwards and to the formation of the fully open
transcription competent RPo2 complex. Also stopped-
flow spectrofluorimetric investigations of the kinetics
of RPo formation at a synthetic promoter bearing con-
sensus )10 and )35 and UP elements have indicated
that DNA opening is not affected by Mg
2+
ions [38].
All these observations allow us to conclude that, in
plasmid-contained Pa and kP
R
promoters, no stable
melting intermediates can be detected by perturbing

the reaction with temperature and Mg
2+
. This point
of view finds support in the recent single-molecule
DNA manipulation experiments [9] on promoter
unwinding by RNAP, which showed that, in this pro-
cess, there are no intermediates with lifetimes longer
than % 1s.
The normal temperature-dependence of the thymine
oxidation reaction demonstrated in RPo calls into
question the results of earlier footprinting investiga-
tions using permanganate and ⁄ or other chemical
probes, in which temperature was used to visualize the
allegedly stepwise opening of stable transcription com-
plexes under the assumption that the underlying chemi-
cal reaction is temperature-independent [10,24–28].
Investigations of thymine oxidation in Thd and its
nucleotides in solution have unequivocally demonstra-
ted that the mechanism underlying the Mg effect on the
reaction in nucleotides is electrostatic and consists of
screening the negative phosphate charges by Mg
2+
,
thereby reducing the electrostatic barrier to diffusion
of MnO
4
–
anions towards the reactive double bond of
thymine. Thus the differences in the rate constants of
thymine oxidation in the absence of Mg

2+
between
nucleotides bearing variously charged and sized phos-
phate groups reflect the differences in the extent of the
electrostatic barrier to MnO
4
–
diffusion, whereas those
extrapolated to the saturating Mg
2+
concentration
point to the differences in the extent of the steric barri-
ers to this process. The general similarity of the kinetic
characteristics of the oxidation reaction in the nucleo-
tide and RPo–kP
R
and RPo–Pa systems allows analog-
ous interpretation of the observed differences in the
rate constants for individual thymines in RPo in terms
of the electrostatic and steric barriers to MnO
4
–
diffu-
sion. Consequently, the observed differences between
the rate constants of thymines variously located in the
separated template and nontemplate promoter strands
can be attributed to the position-dependent extent of
the electrostatic and steric barriers to the diffusion of
the oxidant to the reactive C(5)¼C(6) thymine double
bond. From this perspective, it was worth comparing

the reactivity of thymines in the RPo–kP
R
and RPo–Pa
complexes determined under the same experimental
conditions. This comparison is shown in Fig. 7 as col-
umn plots of
ox
k and
ox
k
Mg
vs. position of the thymine
Fig. 7. Comparison of
ox
k
i
values (37 °C) for thymine oxidation in
RPo–kP
R
and RPo–Pa complexes. RPo–kP
R
, shadowed columns;
RPo–Pa [30], open columns: in the absence of added MgCl
2
(A), in
the presence of 10 m
M MgCl
2
(B), and the Mg effect (C).
T. Łozin

´
ski and K. L. Wierzchowski Open complex at kP
R
promoter
FEBS Journal 272 (2005) 2838–2853 ª 2005 FEBS 2847
in the promoter sequence in relation to the transcrip-
tion start point. It is evident that the general pattern of
relative thymine reactivity in the two complexes is
similar in both the absence and presence of 10 mm
Mg
2+
. Only the absolute values of
ox
k and
ox
k
Mg
for
thymines in RPo–kP
R
are two to threefold higher than
those for the corresponding bases in the RPo–Pa com-
plex. This also applies to T)6, T)7 and T)10, the reac-
tivity of which in the latter complex was too weak to
permit quantitative evaluation of corresponding DNA
bands in the footprints [30]. This seems to be an
important observation, implying higher internal
dynamics of the RPo–kP
R
complex, associated with a

larger flux of MnO
4
–
to the reaction sites. The magni-
tude of the Mg effect,
ox
k
i
,
Mg
⁄
ox
k
i
, however, is compar-
able in the two complexes (Fig. 7C), which indicates
that, in spite of some differences between the sequences
of these promoters in the bubble region (Fig. 1), the
separated DNA strands in both complexes remain in
the same protein environment. The only base that
exhibits different reactivity in the two complexes is
T)9; in RPo–kP
R
in the absence of Mg
2+
its reactivity
is severalfold lower than that of the adjacent T)8 and
apparently insensitive to temperature, whereas in RPo–
Pa the two bases exhibit similar reactivity. However, in
both complexes, the reactivity of these bases increases

similarly in the presence of Mg
2+
. Although in both
promoters T)9 occurs in the same 3¢TAT()9) TA5¢
sequence context, at the crucial )12 position, at which
the separation of DNA strands is believed to be nucle-
ated [1,11,12], they bear different base pairs: GC in
kP
R
and TA in Pa. Thus the GC pair at this position
of the )10 element apparently influences the inter-
actions of T)9 with the protein environment.
The similarity, within an order of magnitude, of
ox
k
values for thymine oxidation in free nucleotides and in
RPo–kP
R
and RPo–Pa indicates generally high struc-
tural dynamics of the bubble region in open transcrip-
tion complexes. The higher dynamics of RPo–kP
R
than that of RPo–Pa can be related to the fact that Pa
bears canonical )10 and )35 recognition hexamers,
whereas in kP
R
they are mutated at positions )12 and
)30, respectively (Fig. 1). Furthermore, kP
R
bears

from )14 to )18 position a row of five GC base pairs,
whereas the whole 17-bp spacer of Pa is made up
solely of AT base pairs. Owing to the presence of a
number of TA steps, the latter is expected to be highly
flexible [39], allowing a better fit between the recogni-
tion hexamers and RNAP.
The greatly varied reactivity of thymines in RPo can
be confronted with the current structural model of the
open transcription complex [11,12]. According to this
model, the two DNA strands separate and take differ-
ent paths beginning at )11 bp, the upstream edge of
the transcription bubble. The )10 nontemplate strand
element from position )11 to )7 crosses the r
2
domain, where it can interact with the exposed aroma-
tic residues of r region 2.3 (corresponding to Tyr425,
Tyr430 and Trp433 of E. coli r
70
). The low reactivity
of T)6, T)7 and T)10 is thus fully consistent with the
proposed stacking interactions between nucleic acid
bases and the aromatic amino-acid side chains in this
region [1,17,40,41]; in a stacked ‘sandwich’ state, thy-
mine can be protected from oxidation by MnO
4
–
either
sterically or by competitive oxidation of its stacked
aromatic partner, or both. The low value of the Mg
effect points to dielectric shielding of these bases. The

downstream part of the nontemplate strand, positions
from )5to)2, is held in a % 10 A
˚
wide tunnel formed
by two lobes of the b (b1 and b2) subunit and the r
2
domain, the 2.1 and 2.3 regions of which provide a
strip of positively charged residues interacting with
DNA phosphates [12]. Again, the very high reactivity
of T)3 and T)4 in RPo–kP
R
and of T)2, T)3 and
T)4 in RPo–Pa, taken together with the moderate Mg
effect, point to partial compensation of the negative
electrostatic potential of DNA phosphates by posi-
tively charged amino acids and to a low steric barrier
to the diffusion of permanganate ions to the thymines
in the wide protein tunnel, in excellent agreement with
the model. The template DNA strand from )11 is
diverted through a % 12 A
˚
tunnel completely enclosed
on all sides by parts of r
2,
r
3
, b1, and the so called
b¢ lid and b¢ rudder [11]. The model predicts that uni-
versally conserved basic amino acids of r regions 2.4
and 3.0 (corresponding to Arg436, Lys462, Arg465

and Arg468 of E. coli r
70
) exposed at the entrance of
the tunnel, and those of region 2.2 on its wall, may
contribute to the electrostatic potential responsible for
stabilization of the template strand in the tunnel. The
high reactivity of T)11 and T)8, connected with a
moderate Mg effect, is thus in good agreement with
the model. The path of the template strand then takes
it past the r
2
– r
3
linker and elements of the b subunit
that make up the back wall of the RNAP active-site
channel, until it passes near the active site [11]. The
very low reactivity of the thymines close to the cata-
lytic center region in the absence of Mg
2+
associated
with the large Mg effect can be interpreted as being
due, for the most part, to a high electrostatic barrier
to diffusion associated with the negative charges of
both DNA phosphates and the carboxylic groups of
three aspartates of the NADFDGD motif involved in
the binding of catalytic Mg
2+
ions [23,42]. Chelation
of these ions to the carboxylic groups may, of course,
induce some local conformational changes around the

Open complex at kP
R
promoter T. Łozin
´
ski and K. L. Wierzchowski
2848 FEBS Journal 272 (2005) 2838–2853 ª 2005 FEBS
transcription start point, also increasing steric accessi-
bility of DNA bases to MnO
4
–
.
In the light of the above considerations, the
Mg-modulated reactivity of thymines reflects variation
in the effective electrostatic potential along the RNAP
channels holding the template and nontemplate DNA
strands in the open complex.
The greater Mg effect on
ox
k
i
in RPo–kP
R
at low
temperatures, causing an apparent decrease in the cal-
culated
ox
E
a
values, deserves a short comment. It was
not observed for the thymidine nucleotides and hence

cannot be ascribed to a higher extent of diffusive bind-
ing of Mg
2+
to DNA phosphates, which would lead
to an increase in K
scr
. However, K
ass
for specific chela-
tion of these ions by carboxylate groups can be expec-
ted to increase with a fall in temperature. Besides
aspartates in the catalytic center, these groups may
occur in the protein channels in proximity to the strips
of positively charged basic amino-acid residues, with
which they might form salt bridges before isomeriza-
tion of the complex and DNA strand separation [32].
In our recent work on KMnO
4
oxidation of RPo–Pa
[29], we showed that Mg
2+
ions exert an opposite,
protective, effect on the protein part of the complex,
by slowing down oxidation reactions of some amino
acids in RNAP, strongly competitive with thymine oxi-
dation in the melted DNA region. In the presence of
10 mm Mg
2+
, the rates of RPo inactivation at 0 °C
and dissociation of strongly damaged RPo,ox com-

plexes into components at 35 °C are twofold smaller.
At lower temperatures, both the expected higher occu-
pancy of Mg-chelating sites and the smaller amplitude
of breathing fluctuations in RNAP would diminish dif-
fusion of MnO
4
–
to vulnerable sites. Simultaneous
enhancement by Mg
2+
of the oxidation reaction of
thymines makes it more competitive, leading to higher
ox
k
i,Mg
⁄
ox
k
i
ratios. Thus, the inverse relationship
between
ox
k
i,Mg
⁄
ox
k
i
and temperature, causing an
apparent decrease in the corresponding activation ener-

gies of thymine oxidation, should not be considered in
connection with the
ox
E
a
values determined in the
absence of Mg
2+
.
Concluding remarks
Our present and earlier studies [29,30] on permangan-
ate oxidation of RPo clearly show that careful applica-
tion of this oxidant as a probe for melted DNA regions
in DNA–protein complexes may yield precise informa-
tion on both the position-dependent inherent reactivity
of vulnerable sites in DNA and the structural dynamics
of the complexes. Moreover, when combined with
modulation of the electrostatic barrier to diffusion of
the anionic oxidant by Mg
2+
ions, this method may
also provide information on the distribution of the
electrostatic potential along the single DNA strands.
Experimental procedures
RNA polymerase
RNA polymerase (EC 2.7.7.6) was prepared from E. coli
C600 strain as described by Burgess & Jendrisak [43] except
that Sephacryl S300 was used instead of Bio-Gel A5m.
The enzyme was stored in buffer S (50% glycerol, 100 mm
NaCl, 10 mm Tris ⁄ HCl, pH 7.9, 0.1 mm dithiothreitol) and

its activity was estimated as described previously [44] at
% 50%.
Promoters and DNA primers
Promoter Pa, containing E. coli consensus )35 and )10
hexamers separated by a 17-bp spacer, was synthesized,
cloned into pDS3 plasmid and purified as described previ-
ously [45]. The kP
R
promoter (an 80 bp DNA fragment
from position )59 to +21 with respect to the transcription
start site) was obtained by PCR using lambda phage DNA
as a template and appropriately designed primers with
overhanging ends encoding XhoI and EcoRI sequences. The
PCR product was cloned into the pDS3 plasmid, and the
kP
R
-containing fragment purified as for Pa.
DNA primers used in footprinting of RPo, Pr(nt) – com-
plementary to the nontemplate DNA strand (from +95 to
+116 with respect to the transcription start site of kP
R
),
and Pr(t) – complementary to the template DNA strand
(from )159 to )138) – were synthesized by the solid-phase
phosphoramidite method, purified by denaturing PAGE
followed by DEAE-Sephacel column chromatography and
ethanol precipitation. The 5¢ ends of the primers were phos-
phorylated with a twofold molar excess of [
32
P]ATP[cP]

with polynucleotide kinase. All chemicals were of molecular
biology or reagent grade.
KMnO
4
oxidation
Thymidine and its nucleotides
The reaction was started by adding 0.1 mL freshly prepared
KMnO
4
solution of the desired concentration to 0.9 mL of
substrate solution, 0.111 or 0.055 mm in Thd, TpT, TMP
or TTP, in PSB buffer (1.11 mm potassium phosphate,
pH 7.2, 111 mm KCl) at a selected [MgCl
2
] (0–111 mm).
The final [KMnO
4
] of 0.625–10 m m was always high
enough to ensure pseudo-first-order conditions for the oxi-
dation reaction. The reaction was terminated by adding
0.2 mL 0.2 m NaHSO
3
at various times (30–120 s)
depending on the temperature (1–37 °C) at which it was
carried out. Afterwards the solutions were spun to remove
T. Łozin
´
ski and K. L. Wierzchowski Open complex at kP
R
promoter

FEBS Journal 272 (2005) 2838–2853 ª 2005 FEBS 2849
traces of precipitated MnO
2
, and the decrease in A
267
, pro-
portional to the oxidative conversion of thymine into its di-
glycol, was measured spectrophotometrically.
RPo at the Pa promoter
Oxidation of thymines in RPo at the Pa promoter and detec-
tion of the oxidation products by primer extension reaction
were performed as described previously [29,30]. The com-
plex, formed by incubation for 15 min at 37 °C of 50-lL
samples of the reaction mixture containing 1 pmol pDS3
plasmid and 2 pmol RNAP in TB buffer (25 mm Tris ⁄ HCl,
pH 7.0, 100 mm KCl, 0.2 mm EDTA), was oxidized with a
low single-hit oxidant dose, 0.01 or 0.02 m · s, at a number
of MgCl
2
concentrations spanning the range 0–40 mm. The
experiments were performed in duplicate. The total intensity
of the group of DNA bands corresponding to the nontem-
plate thymines from T)4 to T+3 was normalized to the
intensity of the whole lane; both intensities were determined
by volume integration using the imagequant software.
RPo at the kP
R
promoter
Footprinting reactions of RPo at kP
R

were performed as a
function of single-hit oxidant dose as described for RPo at
the Pa promoter [29,30]. RPo was formed as described in a
recent kinetic study [6] at three selected temperatures of
13 °C, 25 °C and 37 °C using incubation times in TB buffer
of 120, 60 and 30 min, respectively, and fourfold excess of
RNAP.
KMnO
4
of the desired concentration (freshly diluted from
0.1 m stock) was added, and the reaction mixture incubated
for precisely 1 min at the temperature of the open complex
formation. Note that [KMnO
4
] and time of the reaction
were shown to be equivalent variables of the kinetics [29].
The following oxidant doses (m · s) were applied: at 37 °C
)0.004, 0.008, 0.012, 0.016, 0.02; at 25 °C )0.01, 0.02, 0.03;
at 13 °C )0.01, 0.02, 0.03 and 0.04. The nominal doses were
corrected for the redox capacity of the buffer [29]. Oxidized
DNA bases were detected by the primer extension reaction
using
32
P-end-labeled primers. Purified products of the reac-
tion were resolved (in duplicate) on 6% polyacrylamide
sequencing gels in TBE buffer (0.089 m Tris ⁄ borate, pH 8.3,
and 2 mm EDTA). Dried gels were exposed to storage phos-
phor screens (Molecular Dynamics, Sunnyvale, CA, USA).
From each pair of gels, the one exhibiting a better resolu-
tion of bands was selected for quantitative analysis. All

footprinting experiments were duplicated or triplicated.
Phosphorimager analysis and quantification
of band intensities
Images of footprints were obtained with the use of a
Molecular Dynamics Phosphorimager. Integrated intensities
of bands (or groups of bands) and their intensity profiles
along gel lanes were obtained with the help of the image-
quant software, by volume integration and area integration
of quadrilateral contours encompassing these bands,
respectively. For area integration, the lowest intensity point
in the graph was used as the horizontal baseline (back-
ground). In volume integration, local background was used
for bands corresponding to promoter bubble DNA frag-
ments, whereas for the whole lanes that of the gel without
any radioactivity was used (outside the lanes). These data
were analyzed further using originä software (MicroCal
Software, Northampton, MA, USA).
Fractions
32P
f
i
of DNA fragments corresponding to oxid-
ized T+1, T+2 and the sum of T)7 and T)10, character-
ized by well-separated band profiles, were determined from
integrated band intensities within respective quadrilateral
contours: (+2 … )1) (+3 … )1) and ()6 … )11), and
normalized to that of the whole line.
Fractions
32P
f

i
of DNA fragments characterized by over-
lapping band profiles, corresponding to T)11, T)9, T)8
(within the )6 … )13 quadrilateral contour) and to T)4
and T)3 (within the )2 … )5 quadrilateral contour), were
calculated after deconvolution of the integrated and nor-
malized contour profiles into peaks, with use of a Lorentz
distribution function, to obtain the actual band partition.
The primer extension reaction stops at a base preceding
the oxidized thymine when the glycol form of the latter is
hydrolyzed to a urea derivative during alkaline denatura-
tion [46], which is clearly seen as band doubling for T)11
and T+2. Therefore, the measured contribution of these
forms to the integrated intensity of these two bases was
used in the calculation of the contribution of individual
bases to the integrated intensity of overlapping DNA
bands.
Determination of the promoter occupancy
Products of the Klenow reaction can also include DNA
fragments originating from unmodified plasmid molecules
and ⁄ or those complexed with RNAP at sites other than the
promoter. Therefore, for proper interpretation of the sin-
gle-exponential dependence of
32P
f
i
on the oxidant dose –
32P
f
i

¼ Fo (1-exp[–k
i
x]) – one has to know the actual frac-
tion Fo of promoter DNA occupied by RNAP in the form
of RPo. Values of Fo were determined by simultaneous fit-
ting of the above equation to the experimental
32P
f
i
data
(four sets for each strand at a given temperature, in the
absence and presence of 10 mm MgCl
2
), together with the
fraction of unmodified DNA:
32P
f
uDNA
¼ 1–
32P
f
bubble
,
where
32P
f
bubble
¼ S
32P
f

i
. For the more reactive template
strand, footprinted in the presence of 10 mm MgCl
2
, the
following Fo values were obtained: 0.91 ± 0.04,
0.90 ± 0.01 and 0.81 ± 0.05, at 37 °C, 25 °C and 13 °C,
respectively. As the promoter occupancy should not be
Open complex at kP
R
promoter T. Łozin
´
ski and K. L. Wierzchowski
2850 FEBS Journal 272 (2005) 2838–2853 ª 2005 FEBS
smaller in a lower salt buffer, the same Fo values were
assumed to represent fractions of RPo in footprinting reac-
tions carried out in both the presence and absence of
10 mm MgCl
2
. Consequently, from the experimental frac-
tions corrected for the promoter occupancy,
32P
f
i
⁄ Fo ¼
RPo
f
i
, the actual fractions of oxidized thymines in RPo were
calculated as follows:

ox
f
i
¼
RPo
f
i
=
X
i
i¼1
RPo
f
i
þ
RPo
f
uDNA
!
where
RPo
f
uDNA
¼ 1 À
X
i
i¼1
RPo
f
i

is the fraction of unmodified DNA in RPo; i numbering
starts from the top of gel lanes.
Acknowledgements
The authors thank Dr Krystyna Bolewska and Teresa
Rak for preparation of RNA polymerase of excellent
quality.
References
1 Helmann JD & deHaseth PL (1999) Protein–nucleic
acid interactions during open complex formation inves-
tigated by systematic alteration of protein and DNA
binding partners. Biochemistry 38, 5959–5967.
2 Roe J-H, Burgess RR & Record MT Jr (1984) Kinetics
and mechanism of the interaction of Escherichia coli
RNA polymerase with the kP
R
promoter. J Mol Biol
176, 495–521.
3 Roe J-H, Burgess RR & Record MT Jr (1985) Tem-
perature dependence of the rate constants of the Escher-
ichia coli RNA polymerase–kP
R
promoter interactions.
Assignment of the kinetic steps corresponding to protein
conformational change and DNA opening. J Mol Biol
184, 441–453.
4 Suh WC, Leirmo S & Record MT Jr (1992) Role of
Mg
2+
in the mechanism of formation and dissociation
of open complexes between Escherichia coli RNA

polymerase and the kP
R
promoter: kinetic evidence for
a second open complex requiring Mg
2+
. Biochemistry
31, 7815–7825.
5 Tsodikov OV & Record MT Jr (1999) General method
of analysis of kinetic equations for multistep reversible
mechanism in single-exponential regime: applications to
kinetics of open complex formation between Er
70
RNA
polymerase and kP
R
promoter DNA. Biophys J 78,
1320–1329.
6 Saecker RM, McQuade LM, Schlax PE Jr, Capp MW
& Record MT Jr (2002) Kinetic studies and structural
models of the association of E. coli RNA polymerase
with the kP
R
promoter: large scale conformational
changes in forming the kinetically significant intermedi-
ates. J Mol Biol 319, 649–671.
7 Buc H & McClure WR (1985) Kinetics of open complex
formation between Escherichia coli RNA polymerase
and the lacUV5 promoter. Evidence for a sequential
mechanism involving three steps. Biochemistry 24, 2712–
2723.

8 Rosenberg S, Kadesch TR & Chamberlin MJ (1982)
Binding of Escherichia coli RNA polymerase holo-
enzyme to bacteriophage T7 DNA. Measurements of
the rate of open complex formation at T7 promoter A1.
J Mol Biol 155, 31–51.
9 Revyakin A, Ebright RH & Strick TR (2004) Pro-
moter unwinding and promoter clearance by RNA
polymerase: detection by single-molecule DNA nano-
manipulation. Proc Natl Acad Sci USA 101, 4776–
4780.
10 Suh WC, Ross W & Record MT Jr (1993) Two open
complexes and a requirement for Mg
2+
to open the
kP
R
transcription start site. Science 259, 358–361.
11 Murakami KS, Masuda S, Campbell EA, Muzzin O &
Darst SA (2002) Structural basis of transcription initia-
tion: an RNA polymerase holoenzyme–DNA complex.
Science 296, 1285–1290.
12 Mekler V, Korthkonija E, Mukhopadhyay J, Knight J,
Revyakin A, Ebright YW, Levy R & Ebright RH
(2002) Structural organization of bacterial RNA poly-
merase holoenzyme and the RNA polymerase-promoter
open complex. Cell 108, 599–614.
13 Vassylyev DG, Sekine S-I, Laptenko O, Lee J,
Vassylyeva MN, Burukhov S & Yokoyama S (2002)
Crystal structure of bacterial polymerase holoenzyme at
2.6 A

˚
resolution. Nature 417, 712–719.
14 Murakami KS, Masuda S & Darst SA (2002) Structural
basis of transcription initiation: RNA polymerase
holoenzyme at 4 A
˚
resolution. Science 296, 1280–1284.
15 Severinov K & Darst SA (1997) A mutant RNA pol-
merase that forms unusual open promoter complexes.
Proc Natl Acad Sci USA 94, 13481–13486.
16 Nechaev S, Chlenov M & Severinov K (2000) Dissec-
tion of two hallmarks of the promoter open complex by
mutation in an RNA polymerase core subunit. J Biol
Chem 275, 25516–25522.
17 Young BA, Anthony LC, Gruber TM, Arthur TM,
Heyduk E, Lu CZ, Sharp MM, Heyduk T, Burgess RR
& Gross CA (2001) A coiled-coil from the RNA poly-
merase beta¢ subunit allosterically induces selective non-
template strand binding by sigma (70). Cell 105, 935–
944.
18 Young BA, Gruber TM & Gross CA (2004) Minimal
machinery of RNA polymerase holoenzyme sufficient
for promoter melting. Science 303, 1382–1384.
19 Krakow JS, Rhodes G & Jovin TM (1976) RNA poly-
merase: catalytic mechanisms and inhibitors. In RNA
T. Łozin
´
ski and K. L. Wierzchowski Open complex at kP
R
promoter

FEBS Journal 272 (2005) 2838–2853 ª 2005 FEBS 2851
Polymerase (Losick R & Chamberlin M, eds), pp. 127–
157. Cold Spring Harbor Laboratory Press, Plainview,
NY.
20 Koren R & Mildwan AS (1977) Magnetic resonance
and kinetic studies of the role of the divalent cation
activator of RNA polymerase from Escherichia coli.
Biochemistry 16, 241–249.
21 Szafran
´
ski P, Smagowicz JW & Wierzchowski KL
(1985) Substrate selection by RNA polymerase from
E. coli. The role of ribose and 5¢-triphosphate frag-
ments, and nucleotides interaction. Acta Biochim Pol 32,
329–349.
22 Mustaev A, Kozlov M, Markovtsov V, Zaychikov E,
Denissova L & Goldfarb A (1997) Modular organiza-
tion of the catalytic center of RNA polymerase. Proc
Natl Acad Sci USA 94, 6641–6645.
23 Sosunov V, Sosunova E, Mustaev A, Bass I, Nikiforov
V & Goldfarb A (2003) Unified two-metal mechanism
of RNA synthesis and degradation by RNA polymer-
ase. EMBO J 22, 2234–2244.
24 Craig ML, Suh WC & Record MT Jr (1995) HO and
DNase I probing of Er
70
RNA polymerase–kP
R
open
complexes: Mg

2+
binding and its structural conse-
quences at the transcription start site. Biochemistry 34,
15624–15632.
25 Craig ML, Tsodikov OV, MacQuade KL, Schlax PE,
Saecker RM, Capp MW & Record MT Jr (1998) DNA
footprints of the two kinetically significant intermediates
in formation of an RNA-polymerase-promoter open
complex: evidence that interactions with start site and
downstream DNA induce sequential conformational
changes in polymerase and DNA. J Mol Biol 283, 741–
756.
26 Zaychikov E, Denissova L, Meier T, Gotte M &
Heumann H (1997) Influence of Mg
2+
and temperature
on formation of the transcription bubble. J Biol Chem
272, 2259–2267.
27 Chen Y-F & Helmann JD (1997) DNA melting at the
Bacillus subtilis flagellin promoter nucleates near )10
and expands unidirectionally. J Mol Biol 267, 47–59.
28 Tsodikov OV, Craig ML, Saecker RM & Record MT Jr
(1998) Quantitative analysis of multiple-hit footprinting
studies to characterize DNA conformational changes in
protein-DNA complexes: application to DNA opening
by Er
70
RNA polymerase. J Mol Biol 283, 757–769.
29 Łozin
´

ski T & Wierzchowski KL (2003) Inactivation and
destruction by KMnO
4
of Escherichia coli RNA poly-
merase open transcription complex: recommendations
for footprinting experiments. Anal Biochem 320, 239–
251.
30 Łozin
´
ski T & Wierzchowski KL (2001) Mg
2+
ions do
not induce expansion of the melted DNA region in
the open complex formed by Escherichia coli RNA
polymerase at a cognate synthetic Pa promoter. A
quantitative KMnO
4
footprinting study. Acta Biochim
Pol 48, 495–510.
31 Kolasa IK, Łozin
´
ski T & Wierzchowski KL (2001)
Mg
2+
does not induce isomerization of the open tran-
scription complex of Escherichia coli RNA polymerase
at the model Pa promoter bearing consensus )10 and
–35 hexamers. Acta Biochim Pol 48, 985–994.
32 Saecker RM & Record MT Jr (2002) Protein surface
salt bridges and paths for DNA wrapping. Curr Opin

Struct Biol 12, 311–319.
33 Ł ozin
´
ski T & Wierzchowski KL (2001) Effect of Mg
2+
on kinetics of oxidation of pyrimidines in duplex DNA
by potassium permanganate. Acta Biochim Pol 48, 511–
523.
34 Hayatsu H & Ukita T (1967) The selective degradation
of pyrimidines in nucleic acids by permanganate oxida-
tion. Biophys Biochem Res Commun 29, 556–561.
35 Saenger W (1984) Principles of Nucleic Acid Structure.
Springer Verlag, New York.
36 Poznan
´
ski J, Felczak K, Kulikowski T & Remin M
(2000)
1
H NMR conformational study of antiherpetic
C-substituted 2¢ deoxyuridines: insight into the nature
of structure–activity relationships. Biochem Biophys Res
Commun 272, 64–74.
37 Walaas E (1958) Stability constants of metal com-
plexes with mononucleotides. Acta Chem Scand 12 ,
528–536.
38 Strainic MG Jr, Sullivan JJ, Velevis A & deHaseth PL
(1998) Promoter recognition by Escherichia coli RNA
polymerase: Effects of the UP element on open complex
formation and promoter clearance. Biochemistry 37,
18074–18080.

39 Kolasa IK, Łozin
´
ski T & Wierzchowski KL (2003)
Effects of distortions by A-tracts of promoter B-DNA
spacer region on the kinetics of open complex formation
by Escherichia coli RNA polymerase. Acta Biochim Pol
50, 909–920.
40 Juang Y-L & Helmann JD (1994) A promoter melting
region in the primary sigma factor of Bacillus subtilis.
Identification of functionally important aromatic amino
acids. J Mol Biol 235, 1470–1488.
41 Malhotra A, Severinova E & Darst SA (1996) Crystal
structure of a r
70
subunit fragment from E. coli RNA
polymerase. Cell 87, 127–136.
42 Zaychikov E, Martin E, Denissova L, Kozlov M,
Markovtsov V, Kashlev M, Heumann H, Nikiforov V,
Goldfarb A & Mustaev A (1996) Mapping of catalytic
residues in the RNA polymerase active center. Science
273, 107–109.
43 Burgess RR & Jendrisak JJ (1975) A procedure for the
rapid, large scale purification of Escherichia coli DNA-
dependent RNA polymerase involving polyamine P
precipitation and DNA-cellulose chromatography.
Biochemistry 14, 4634–4638.
Open complex at kP
R
promoter T. Łozin
´

ski and K. L. Wierzchowski
2852 FEBS Journal 272 (2005) 2838–2853 ª 2005 FEBS
44 Chamberlin M, Kingstone R, Gilman M, Wiggs J &
deVera A (1983) Isolation of bacterial and bacterio-
phage RNA polymerases and their use in synthesis of
RNA in vitro. Methods Enzymol 101, 40–568.
45 Łozin
´
ski T, Adrych-Rozek K, Markiewicz WT & Wierz-
chowski KL (1991) Effect of DNA bending in various
regions of a consensus-like Escherichia coli promoter on
its strength in vivo and structure of the open complex
in vitro. Nucleic Acids Res 19, 2947–2953.
46 Ide H, Kow YW & Wallace SS (1985) Thymine
glycols and urea residues in M13 DNA constitute
replicative blocks in vitro. Nucleic Acids Res 13, 8035–
8052.
T. Łozin
´
ski and K. L. Wierzchowski Open complex at kP
R
promoter
FEBS Journal 272 (2005) 2838–2853 ª 2005 FEBS 2853