© 7990 Oxford University Press
Nucleic Acids Research, Vol. 18, No. 15 4369
The double role of methyl donor and allosteric effector of
S-adenosyl-methionine for Dam methylase of E. coli
Agnes Bergerat and Wilhelm Guschlbauer
Service de Biochimie, Bat. 142, Departement de Biologie, Centre d'Etudes Nucleates de Saclay,
F-91191 Gif-sur-Yvette Cedex, France
Received May 26, 1990; Revised and Accepted July 3, 1990
ABSTRACT
The turnover of DNA-adenine-methylase of E. coli
strongly decreases when the temperature is lowered.
This has allowed us to study the binding of Dam
methylase on 14 bp DNA fragments at 0°C by gel
retardation in the presence of Ado-Met, but without
methylation taking place. The enzyme can bind nonspecific DNA with low affinity. Binding to the specific
sequence occurs in the absence of S-adenosylmethionine (Ado-Met), but is activated by the presence
of the methyl donor. The two competitive inhibitors of
Ado-Met, sinefungin and S-adenosyl-homocysteine,
can neither activate this binding to DNA by themselves,
nor inhibit this activation by Ado-Met. This suggests
that Ado-Met could bind to Dam methylase in two
different environments. In one of them, it could play
the role of an allosteric effector which would reinforce
the affinity of the enzyme for the GATC site. The
analogues can not compete for such binding. In the
other environment Ado-Met would be in the catalytic
site and could be exchanged by its analogues. We have
also visualized conformational changes in Dam
methylase induced by the simultaneous binding of AdoMet and the specific target sequence of the enzyme,
by an anomaly of migration and partial resistance to
proteolytic treatment of the ternary complex AdoMet/Dam methylase/GATC.
site (5). Methylation of the recognition sequence is modulated
by the nature of the three base pairs flanking both sides of the
site (6). The interaction of the enzyme with its site must therefore
be asymmetric.
The mechanism of Dam methylase could be decomposed into
three steps: specific recognition of the GATC sites; binding of
the cofactor, Ado-Met; methyl transfer from Ado-Met to the
adenine of the target site. These steps could, however, overlap.
We do not know whether the sequence specificity of Dam
methylase is manifested at the level of DNA binding or of
catalysis. The cofactor could also influence this specificity.
To elucidate these points we have performed gel retardation
binding experiments of Dam methylase on oligonucleotides, with
or without the recognition sequence. In contrast to the restriction
endonucleases, it is impossible to isolate quantitatively on
nitrocellulose filters specific complexes between methylases and
DNA in the absence of Ado-Met (7). A role of the cofactor in
this specific recognition, similar to that of the type I restrictionmodification systems (8), could be proposed. In the presence of
Ado-Met, however, methylation proceeds rapidly and no stable
complexes can be isolated. To study the influence of Ado-Met
in the absence of catalysis we have worked at low temperature
where the turnover of Dam methylase is insignificant. We have
also studied the influence on the specificity of Dam methylase
of two analogues which are competitive inhibitors for Ado-Met,
adenosyl-ornithine (sinefungin) and S-adenosyl-homocysteine
(Ado-Hcy).
INTRODUCTION
DNA-adenine-methylase (Dam methylase) of Escherichia coli
recognizes specifically the sequence GATC in DNA and transfers
a methyl group from S-adenosyl-methionine (Ado-Met) to the
amino group of the adenine (1). This enzyme methylates DNA
after replication with a lag time (2), leaving the GATC sites of
the DNA temporarily hemimethylated. The state of methylation
(un-, hemi- or bimethylated) of the GATC site is a signal
recognized by many systems in the cell: post-replicative mismatch
repair, replication, transcription of certain genes, transposition,
and segregation of the chromosome (3,4). Although the biological
role of Dam methylase is partially elucidated, its mechanism of
action is unknown. It constitutes a good model to study sequence
specific DNA-protein interactions.
Dam methylase is a protein of 32 kD. It acts as a monomer
and transfers one methyl group per binding event onto the GATC
MATERIALS AND METHODS
The oligonucleotides were prepared on an Applied Biosystem
model 38 DNA synthesizer using phosphoamidite solid phase
chemistry (Table 1). All oligonucleotides were asymmetric,
containing two A.T pairs on one side of the GATC site and G.C
pairs on the other as it had been shown (6) that with such
asymmetric sequences the site is preferentially methylated. A 10
bp long oligonucleotide duplex had been synthesized by A. Guy
et R. Teoule (CEN-Grenoble). Other oligonucleotides (12 bp and
14 bp) were purified on a reversed-phase C1/C8 column
(ProRPCHCR/2 from Pharmacia) (9). The pBR322 DNA
fragment (Nrul-Sall) used in this work was 321 bp long. It was
purified by polyacrylamide gel electrophoresis. DNA fragments
were 5' end labeled by using T4 polynucleotide kinase from
4370 Nucleic Acids Research, Vol. 18, No. 15
Tablel: Oligonucleotide sequences used.
10SP
51 GCGATCATGG 3'
31 CGCTAGTACC 51
12SP
5'GGCGGATCAACG 31
3'CCGCCTAGTTGC 5'
14SP
5'GCGCGATCATGGCG 31
3'CGCGCTAGTACCGC 51
14NS
5'GCGCGTACATGGCG 51
3'CGCGCATGTACCGC 3'
SP: DNA fragments carrying the GATC sequence specific for Dam methylase;
NS: DNA fragment without this specific sequence
NH,
HO
OH
NHj,
OH
H00
HO
OH
Figure 1: Cofactors used: a) S-Adenosyl-methionine (Ado-Met), b) adenosylomithine (sinefungin), c) S-adenosyl-homocysteine (Ado-Hcy)
Boehringer Mannheim and ( T - 3 2 P ) ATP (3000 Ci/mmole) from
Amersham.
Plasmid pBR322 DNA lacking the Dam methylase modification
was isolated from the GM33 strain of E. coli (which has a
mutation at the dam-3 locus), obtained from Prof. P. Modrich
(Duke University) and purified by the method described in (10),
followed by CsCl/ethidium bromide equilibrium density
centrifugation.
Sinefungin and S-isobutyl-adenosine (SIBA) were generous
gifts from Dr. M. Gero (C.N.R.S. Gif-sur-Yvette, France).
Adenosyl-methionine (Ado-Met) and adenosyl-homocysteine
(Ado-Hcy) (Figure 1) were purchased from Sigma. Tritiated Ado-
Met (15 Ci/mmole) was purchased from Amersham. Purity of
Ado-Met was checked by HPLC (11). Homogenous Dam
methylase was prepared from E. coli strain GM2290/pTP166,
obtained from Prof. M. Marinus (University of Massachusetts),
as described previously (12). Proteinase K was purchased from
Boehringer Mannheim. GF/C glass fiber and DE81 filters were
purchased from Whatman.
Protein and DNA concentration determination
Protein content was determined by the method of Bradford (13),
using bovine serum albumin as standard. A molecular mass of
32 kD for Dam methylase was employed for determining molar
yield. DNA concentrations' were determined spectrophotometrically. The e26o f° r t n e t w o strands of each
oligonucleotide was calculated by the sum of the e260 °f e a c h °f
the bases of the strands in the absence of stacking at 95 °C.
Dam methylase assays
The Dam methylase assay (1) measures the transfer of
(3H)methyl groups from Ado-Met to calf thymus DNA. One
unit of Dam methylase is defined as the amount of enzyme which
catalyzes the conversion of one pmole of (^methyl groups into
a form which binds to DE81 filters in 30 min. at 37°C.
Steady state kinetics
The reaction mixture (50 /tl) contained Tris-HCl, pH8, 150 mM;
dithiothreitol, 2 mM; bovine serum albumin, 400 /tg/ml; EDTA,
10 mM; pBR322 DNA, 9 nM (200 nM in Dam recognition sites)
and concentrations of Ado-Met from 1.67—13.3 jtM. Dam
methylase concentrations and incubation times were adjusted to
obtain from 1 to 10% of total GATC sites methylated at the end
of the reaction. Dam methylase was diluted in the reaction buffer.
Aliquots (20 /tl) were removed, mixed with 10 /tl of non
radioactive Ado-Met (4 mM) in HC1 (0. IN), spotted onto GF/C
filters. DNA was precipitated on the filter by immersion in 5
ml trichloracetic acid (10%), washed four times with 5 ml
trichloracetic acid (10%) and twice with ethanol, dried and
counted in a scintillator. The rate of methyl transferred to the
DNA, vj (in moles of methyl groups transferred per minute per
mole of enzyme) was calculated from vi = (cpmb-cpmo)/(p.t.e),
where cpmb is the radioactivity bound to the GF/C filters, cpm,,
the radioactivity at the zero time, p the total radioactivity per
mole of methyl group, t the time of incubation in minutes, e the
number of pmoles of Dam methylase per aliquot.
Inhibition constants of Ado-Met analogues
Ado-Met and analogues (Ado-Hcy and sinefungin) at indicated
concentrations were preincubated with the reaction medium,
except Dam methylase, during 2 minutes at various temperatures.
Reactions were started by adding the enzyme and stopped as
described above.
Gel retardation assays
DNA-protein complexes were formed in 20 jtl of Dam assay
buffer plus 5% glycerol. The different labeled oligonucleotides,
Dam methylase and cofactors were mixed at 0°C at indicated
concentrations. After an incubation for 2 minutes in ice, samples
were loaded on polyacrylamide gels prepared as previously
described (14). The gels were autoradiographed after drying on
preflashed Kodak-X OMATS films at -70°C with Cronex HI
plus intensifying screens.
Nucleic Acids Research, Vol. 18, No. 15 4371
«
-1 -
o
-4
3,3
3,4
3,5
3,6
3,7
1000/T(K-1 )
0,0
0,2
0,4
0,6
0,8
1,0
Figure 2: Temperature dependence of the turnover of the Dam methylase using
pBR322 DNA as substrate. Vmax of Dam methylase were determined at 30°C,
22°C, 15°C and 8°C, in the presence of 200 nM of GATC sites and with AdoMet concentrations from 1.67 to 13 /tM as described in Materials and Methods.
The value corresponding to the solid symbol is from ref. (5) on ColEl DNA.
Limited proteolysis
Limited proteolysis was performed on Dam methylase already
associated in complexes with the MSP or 14NS DNA fragment
in presence or in absence of Ado-Met. The complexes were
formed as described above and 2 /xl of various concentrations
of proteinase K were added. The incubation was carried out in
ice for 15 minutes. The samples were immediately loaded on
a 10% polyacrylamide gel.
RESULTS
Temperature dependence of some thermodynamic and kinetic
parameters of methylation by Dam methylase
To work in the presence of the cofactor but in absence of catalysis
the temperature dependence of the methyl transfer reaction by
Dam methylase was investigated. In the presence of saturating
amounts of pBR322 DNA (200 nM of GATC sites corresponding
to 25 times the apparent Km in ColEl DNA (5)), the initial
velocities were measured for Ado-Met concentrations from 1.67
to 13.3 nM. As expected the velocity of methylation increased
with temperature. From an Arrhenius plot the turnover was
shown to be strongly temperature dependent (Figure 2). The
activation energy of the reaction was estimated to be 93 kJ/mole
which corresponds to a four fold decrease of the rate every 10°C.
At 0°C k(.at is below 0.001 mole of methyl groups per mole of
enzyme per second.
Inhibition constants of the analogues of Ado-Met, sinefungin
and Ado-Hcy, were also determined in the presence of saturating
amounts of DNA. Both analogues showed competitive inhibition
(Figure 3) while SIB A was without effect (not shown). The
inhibition constants of sinefungin and Ado-Hcy as well as the
Km appear similar over the temperature range 8°C to 30°C
(Table 2).
As the turnover of the enzyme even at 30°C (0.06 s"1) is low
and the Km is rather high, the Kj of the Ado-Met and its
analogues could be approximated by their Km and Ki;
respectively, Km = (kd + k,..,,)/!^. The observed upper values of
0,2
0,4
0,6
0,8
Figure 3: Competitive inhibition of methylation by Ado-Hcy and sinefungin. Eadie
plots of the initial rates of methylation (V) with Ado-Met concentrations (S) from
1.67 to 13.3 /iM and different fixed concentrations of Ado-Hcy (top) and sinefungin
(bottom). The reactions were assayed as described in Methods.
association rate constants (kj for enzyme substrate interactions
fall in the range of 106 to 108 s-'M" 1 (15). Taking Ic, = 106,
kj = (Km . k j - k c a = 6 s-'. As keat = 0.06 s~' at 30°C and
is smaller at lower temperature, it is negligible compared to 6
s~'. The Km and the Kj values give therefore the Kj for AdoMet and its analogues for Dam methylase.
Influence of the size of the DNA fragment on the nature of
the complex formed between DNA and Dam methylase
The first binding assays were performed with a 321 bp DNA
fragment from pBR322 with one GATC site. This sequence was
chosen as being long enough to try footprinting experiments in
4372 Nucleic Acids Research, Vol. 18, No. 15
Table 2: Temperature dependence of the K^ of Ado-Met and the Kj of sinefungin
and Ado-Hcy.
Temperature
(in °C)
, of Ado-Met
(in
Kj of Ado-Hcy
(in fiM.)
Kj of sinefungin
(in /iM)
30
22
15
8
6.5 ±0.7
5.6±1
7.0±2
3.0±1.5
7.0±0.5
5.0±2
5.0±2
2.5±2
2.0±0.6
1.4±1
3.0±l
2.5±1
A
1 2 3 4 5 6 7 8 9
B
10 1 2 3 4 5 6 7 8 9
10
Figure 4: Association of Dam methylase with MSP and 14NS DNA fragments
in the absence or the presence of Ado-Met. Panel A: Lanes 1 - 5 : 50 nM of 14SP
DNA fragment in the absence of Ado-Met were incubated at 0°C with Dam
methylase concentrations of 50, 100, 200, 400 and 800 nM. Lanes 6-10: 10
nM of 14SP DNA fragment in the presence of 100 /»M of Ado-Met were incubated
at 0°C with Dam methylase concentrations of 5, 10, 20, 40 and 80 nM. Panel
B: 50 nM of 14NS DNA fragment were incubated at 0°C with Dam methylase
concentrations of 50, 100, 200, 400 and 800 nM Lanes 1-5: in the absence
of Ado-Met . Lanes 6-10: in the presence of 100 /M Ado-Met.
the case of specific binding. This fragment incubated with Dam
methylase gave rise to a shifted pattern on acrylamide gel with
a number of bands increasing with the quantity of enzyme added
(results not shown). These bands corresponded to the probe
shifted by one, two or more molecules of Dam methylase. There
was no accumulation of the first complex (where the probe
complexed just one molecule of enzyme). This indicated that Dam
methylase exhibits affinity for unspecific DNA. The affinity of
Dam methylase for specific and unspecifc DNA can not be greatly
different.
To investigate these potential differences of affinities and the
influence of the cofactor, the next experiments were performed
with DNA fragments short enough that only one binding site for
the Dam methylase was available. The minimum length of DNA
required for binding of Dam methylase was therefore determined.
Dam methylase can methylate (with a low rate, K m = 10 mM at
10°C) a 10 bp oligonucleotide formed by a GATC site surrounded
by four base pairs on one side and two on the other (Table 1).
But no shifts could be visualized with such a DNA fragment on
retardation gels; not even a smear of free DNA was observed.
In contrast, under the same conditions at 0°C a shift was observed
with a 12 bp oligonucleotide composed of a GATC site flanked
by four bp on each side (Table 1). Binding of Dam methylase
on such a DNA fragment was strongly increased by the presence
of Ado-Met.
Association of Dam methylase with the specific and un specific
tetradecamers (14SP and 14NS) in the presence or absence
of Ado-Met
In the following experiments oligonucleotides 14 bp long were
Table 3: Estimation of the K / of the HSP and HNS DNA fragments in the
absence or in the presence of 100 ^M of Ado-Met.
Ado-Met
HSP
HNS
100 ^M
60-70 nM
300-400 nM
1.2-1.6
3-4
ii M
* The Kj's of the 14SP DNA fragment were estimated by evaluation of the Dam
methylase concentrations necessary to complex half of the probe. The Kj's of
the HNS DNA fragment were determined by comparison of the concentrations
of Dam methylase necessary to complex the same quantities of HNS DNA (in
the absence or in the presence of Ado-Met) as HSP DNA in the absence of AdoMet.
Figure 5: Effect of Ado-Met and its analogues on the binding of the Dam methylase
on the 14SP DNA fragment. 10 nM of HSP DNA fragment was incubated at
0°C with 100 nM of Dam methylase with various concentrations of Ado-Met
or analogues or Ado-Met with an analogue. Panel A, lane 1: no enzyme; lanes
2 - 4 : Ado-Hcy 1, 10, 100 /»M; lanes 5 - 7 : sinefungin 1, 10, 100 fiM; lane 8:
no Ado-Met, no analogues. Panel B, lanes 1 - 8 : Ado-Met 100, 200, 400, 600,
800 nM, 2, 4, 10 pM. Panel C, lanes 1 - 2 : Ado-Met 400 nM plus Ado-Hcy
400 nM or 4 jtM; lanes 3 - 4 : Ado-Met 400 nM plus sinefungin 400 nM or 4 ^M.
used. On such DNA fragments, the site concentration is equal
to the duplex concentration as only one molecule of enzyme can
bind an oligonucleotide of 14 bp length. The duplex 14SP (Table
1) contains the GATC site recognized by Dam methylase, flanked
on one side by four and on the other by six base pairs. In the
non-specific duplex 14NS the two central base pairs were inverted
to GTAC.
Figure 4 shows the behaviour of these two 14 bp DNA
fragments incubated with various concentrations of Dam
methylase in the presence or absence of Ado-Met (conditions
specified in the legend). These experiments were performed at
0°C where methyl transfer, checked with tritiated Ado-Met and
unlabelled DNA, was negligible. Even if the fragment did not
contain a specific GATC site the enzyme bound the DNA weakly.
Smears trailing between the free fragment and the complex were
visualized on the gel. Such complexes were partially dissociated
before entrance into the gel and the cage effect was not efficient
to prevent dissociation during migration (Figure 4, panel B). This
is indicative of the weakness of the association between nonspecific DNA and Dam methylase. The presence of Ado-Met
increased the amount of the complexes formed by a factor of
about two but their stability was still very low.
Binding of the enzyme on a DNA fragment carrying the GATC
site was more stable. No smears were observed (Figure 4, panel
A). The Kj was estimated to be 300-400 nM (Table 3).
Comparison of the shifted patterns of the 14SP and 14NS DNA
fragments with increasing concentrations of Dam methylase
allowed us to estimate the affinity of the enzyme for nonspecific
DNA: K d = 3 - 4 /iM in the absence of Ado-Met and Kj
= 1.2-1.6 pM in the presence of 100 /iM of Ado-Met (Table
Nucleic Acids Research, Vol. 18, No. 15 4373
A
1 2 3 4 5 6 7 8 9
Figure 6: Electrophoretic mobility of Dam methylase/14SP DNA fragment
complexes after limited proteolysis by proteinase K. 50 nM of Dam methylase
in the presence of 100 /tM of Ado-Met (panel A) or 500 nM of Dam methylase
in the absence of Ado-Met (Panel B) were incubated with 50 nM of MSP DNA
fragment in 20 fil of reaction medium, for 2 minutes as described in methods
before being submitted to proteolysis. 2 ml of of proteinase K were added to
the previous samples to get a final concentrations of 0, 12.5, 25, 50, 75, 100,
150, 200, and 300 /ig/ml of reaction medium: lanes 2 - 1 0 . Lane 1 was sample
without Dam methylase.
3). Dam methylase appeared to be able to discriminate the GATC
site in the absence of Ado-Met. But the difference of binding
between non-specific DNA and the specific GATC site was rather
small.
In the presence of Ado-Met, the association of Dam methylase
to its specific site was increased by a factor of about five: K^
of the enzyme for the 14SP DNA fragment in presence of 100
mM Ado-Met was about 6 0 - 7 0 nM. All the Kj values
estimated in these experiments should be considered as relative
values and not absolute ones. They are systematically
underevaluated because the half lifes of all these complexes are
not negligible compared to the time necessary to apply the
complexes on the gel and to allow them to enter it.
Influence of the concentration of Ado-Met on the formation
of the ternary complex 14SP DNA/Dam methylase/Ado-Met
The binding of the 14SP DNA fragment by Dam methylase was
measured at Ado-Met concentrations from 100 nM to 10 fiM
(Figure 5, panel A, lane 8 and panel B, lanes 1-8). The Dam
methylase concentration was 100 nM, a ten-fold excess over the
DNA concentration. For 100 nM Ado-Met, the amount of
complex formed was more than twice that found in the absence
of the methyl donor. At 400 nM saturation was reached. For
optimal binding of all the 100 nM of enzyme less than 400 nM
Ado-Met was necessary. To mediate the increase of affinity of
Dam methylase for its target site, Ado-Met seems to bind the
enzyme with a Kd inferior to 100 nM, which means about 100
times less than the binding of the methyl donor to the catalytic
site (K<]= Km = 6 /*M). There was no cofactor, neither in the
gel nor in the migration buffer. Nevertheless, there was no
dissociation during migration of the complex formed between
the 14SP DNA fragment and Dam methylase in the presence of
Ado-Met. This is an indication of a stronger interaction of AdoMet with the enzyme.
Effect of Ado-Met analogues on Dam methylase DNA binding
Ado-Hcy and sinefungin were shown to be competitive inhibitors
of Ado-Met for Dam methylase (Figure 3). Binding of Dam
B
101 2 3 4 5 6 7 8 9 10
Figure 7: Effect of the ionic strength on the binding of Dam methylase to the
MSP DNA fragment in the absence or in the presence of Ado-Met. 10 nM of
MSP DNA fragment was incubated with 100 nM of Dam methylase and no AdoMet (panel A) or 100 mM of Ado-Met (panel B) lane 1, without enzyme; lanes
2 - 1 0 , NaCl concentrations: 5, 20, 50, 75, 100, 150, 200, 300, 500 mM.
methylase on the 14SP DNA fragment was not changed by the
presence of either of these two analogues of Ado-Met (Figure
5, panel A). Concentration ranges used were from 1/10 to 10
times their Kj In the presence of ten times higher concentrations
than that of Ado-Met these two analogues could not suppress the
increase of affinity of Dam methylase for its target site by AdoMet (Figure 5, panel C). This suggests that Ado-Met can bind
Dam methylase in two different ways to provide two different
functions. Firstly, binding with a low Kj, in an environment
where it cannot be displaced by the anologues. Ado-Met would
play the role of an allosteric effector by reinforcing the affinity
of the enzyme for the GATC site. Secondly, to fulfill the methyl
donor role, Ado-Met would be bound less strongly and in a
surrounding where sinefungin as well as Ado-Hcy could compete.
Induced fit to the GATC site of the complex AdoMet/Methylase
The particularity of the complex 14SP DNA fragment/Dam
methylase/Ado-Met was a slightly faster migration than the three
other complexes. Figure 5, panel A lane 8 compared to panel
B clearly shows that in the presence of Ado-Met the complex
14SP DNA fragment/Dam methylase migrated faster than in the
absence of Ado-Met. The migration of the complex formed with
the 14NS DNA fragment was the same in the presence and in
the absence of Ado-Met (Figure 4, panel B lanes 1 - 5 compared
with lanes 6-10) and slower than that of the specific ternary
complex (Figure 4, panel A lanes 5 — 10 compared to panel B).
We also observed that the analogues of Ado-Met did not increase
the affinity of Dam methylase for the GATC site nor form a
ternary complex that migrated faster. The retardation of a DNA
fragment by a protein has been shown to depend on the shape,
the charge and the molecular weight of the protein (16). The
binding of either Ado-Met or the GATC sequence to Dam
methylase did not modify the enzyme conformation, as far as
it can be visualized by gel migration. Apparently, only the
simultaneous binding of Ado-Met and the GATC site induced
a conformational change in the enzyme. The analogues of AdoMet were not able to mediate this effect.
As the difference of migration between the two kinds of
complexes was small we have tried to check for this
conformational change by another method.
Proteolytic treatment of Dam methylase involved in the
complexes with the 14SP DNA fragment, by proteinase K,
exhibits a different pattern in the presence and in the absence
4374 Nucleic Acids Research, Vol. 18, No. 15
of Ado-Met. In the presence of Ado-Met, a small fragment of
the enzyme could be cleaved by proteinase K without disturbing
the affinity for the DNA (Figure 6, panel A). In the absence of
Ado-Met, the binding of Dam methylase to DNA was completely
destroyed by proteolysis (Figure 6, panel B). The formation of
a complex between the 14NS DNA fragment and Dam methylase
was also destroyed by proteolysis treatment in the absence as
well as in the presence of Ado-Met. Folded structural domains
of proteins are relatively resistant to proteolytic cleavage as
compared to the exposed flexible regions that connect them. As
the Dam methylase engaged in the ternary and the binary complex
with the 14SP DNA fragment exhibited a different sensibility
to protease, the organisation of the domain of the two forms of
the enzyme involved in each complex should be different. It
reinforces the hypothesis of a conformational change for Dam
methylase induced by simultaneous binding of Ado-Met and the
GATC site.
Effect of ionic strength on the binding of Dam methylase to
the 14SP and 14NS DNA fragments in the presence or in the
absence of Ado-Met
The nature of the interactions involved in these four complexes
was studied as a function of ionic strength. Except for the ternary
complex, 14SP DNA/Dam methylase/Ado-Met (Figure 7, panel
B), the quantities of the complexes formed with thel4SP DNA
fragment in the absence of Ado-Met (Figure 7, panel A) as well
as those formed with 14NS DNA fragment (not shown) decreased
progressively with increasing NaCl concentration from 5 mM
to 400 mM. The ternary complex with the MSP fragment was
stable up to a NaCl concentration of 75 mM without dissociating.
In the range 75 to 150 mM NaCl the complex began to dissociate.
At 150 mM NaCl about one third of the complex was dissociated.
Above 200 mM the quantity of complex decreased suddenly and
the affinity of Dam methylase for the MSP fragment approached
that observed in the absence of Ado-Met. Morever, migration
of complexes formed at high ionic strength was slower than of
those formed below 200 mM NaCl and appeared to be similar
to the complex formed in the absence of Ado-Met. The interaction
of Dam methylase with non-specific DNA or with the specific
sequence, but in the absence of Ado-Met seems to involve mainly
electrostatic forces which are distroyed by salt. Some hydrophobic
interactions probably exist in the ternary complex.
DISCUSSION
In this paper, we have shown that Dam methylase exhibits a
certain affinity for non-specific DNA. This property was already
encountered in a study of its mechanism of localisation of GATC
sites by binding non-specific DNA and diffusion along the double
helix (6). Independently of the presence of Ado-Met or its
analogues, Dam methylase recognises its specific GATC site.
But the presence of Ado-Met strongly stimulates the binding of
the enzyme to the GATC site and only slightly influences binding
to non-specific DNA.
Results reported in this paper suggest that Ado-Met can bind
to Dam methylase in two different environments. The first and
strongest binding site (with a Kj inferior to 100 nM) is probably
associated with the allosteric role of Ado-Met. The structural
differences between Ado-Met and its analogues (Figure 1) are
the positive charge and/or the methyl group carried by the sulphur
atom. This positive charge was shown to allow the methionine
group of Ado-Met to adopt a cyclic conformation where the
carboxyl group lies near the sulfonium centre. In contrast, the
conformation of homocysteine in Ado-Hcy was found to be
extended, but bent back towards the adenine base (17). The
conformation of sinefungin is expected to be close to that of AdoHcy as this molecule is uncharged. Therefore this part of the
molecule should play a crucial role in the increase of affinity
for the GATC site of Dam methylase induced by Ado-Met.
The allosteric effect of Ado-Met and its association with Dam
methylase is lost at NaCl concentrations above 200 mM. At these
ionic strengths a strong inhibition of methylation was also
observed (5). This inhibition probably arises from the deactivation
of Dam methylase by loss of Ado-Met from the allosteric site,
rather than by dissociation of the enzyme from the DNA which
is more progressive.
The second environment could correspond to the catalytic site
with a higher K,j (about Km = 6 /*M). The analogues can
compete for this site.
Binding of Ado-Met to Dam methylase increases the affinity
of the protein for the specific DNA which could be indicative
of a change in conformation. The formation of the ternary
complex Ado-Met/Dam methylase/GATC site appears to produce
another conformational change, visualized by slightly faster
migration of the complex and partial resistance to proteolysis.
Neither the change in migration nor resistance to proteolysis was
observed for specific DNA in the absence of Ado-Met and
similarly for non-specific DNA with or without Ado-Met. These
two properties were induced by the simultaneous presence of the
methyl donor and of the specific DNA site.
Tritiated Ado-Met has been crosslinked to Dam methylase (18)
by UV irradiation in the absence of DNA. Some other
methyltransferases have already been crosslinked by the same
method (19). Therefore Ado-Met can interact with Dam
methylase in the absence of DNA. Dam methylase does not share
the same mechanism as Hha\ methyltransferase which can not
bind Ado-Met in the absence of its specific sequence (20).
Friedman (21) has observed that fcoRII methyltransferase could
bind its specific sequence in absence of Ado-Met, but that the
presence of the methyl donor or of its analogues stimulated this
binding. The existence of two kinds of binding sites for AdoMet was suggested for this methylase, but the activating site
exhibited a higher Kj than the catalytic one. Morever, analogues
like sinefungine or Ado-Hcy could compete for the activating
site. The BspRl methyltransferase (22) and the EcoRl
methyltransferase (23) exhibit a mechanism similar to that of
EcoRII. The mechanism of Dam methylase would be closer to
that of the Type I restriction-modification systems like EcoK or
EcoB (8). They share the two successive steps of the activation
of DNA binding by Ado-Met and of the conformational change
induced by the interaction with the specific sequence.
The formation of the ternary complex Ado-Met/Dam
methylase/GATC site involves several steps: the binding to nonspecific DNA, the diffusion along the double helix and the fixation
on the GATC site should occur successively. Association of AdoMet with the Dam methylase in the allosteric or in the catalytic
environment could be imposed by the values of their Kj and by
the chronology of the two functions that are fulfilled by the
cofactor. Presence of Ado-Met in the allosteric site modulates
the steps of binding on the DNA but does not induce them. The
association of the ternary complex is probably random, but the
addition of each component tends to reinforce stability and to
favour the formation of the final active complex. Nevertheless,
two phenomena should be discriminated: the allosteric binding
Nucleic Acids Research, Vol. 18, No. 15 4375
of Ado-Met which changes the affinity of Dam methylase for
the DNA substrate and the induced fit to the GATC site of the
complex Ado-Met/enzyme. Many ligand-induced protein
conformational changes have been observed by crystallography
(24). For the sequence specific DNA binding proteins, such as
repressors, which exhibit high association constants for their
specific sites the energy of the hydrogen bonds established
between the base pairs of the specific site and the protein can
not account for these strong affinities. Additional conformational
changes of the protein or DNA are expected to gain the energy
giving rise to these high association constants. Although Dam
methylase does not exhibit such high affinity for its specific
sequence, some conformational change of the protein appears
to be involved in the association of this enzyme with its site.
Allosteric changes in monomeric enzymes have already been
observed. Monomeric hemoglobin from Chironomus thummi
thummi (25) has been shown to exhibit a Bohr effect. The
ribonucleotide reductase from Lactobacillus leichmanii is a
monomer and requires adenosyl-cobalamine for activity. The
enzyme possesses a regulatory site. Binding of
deoxyribonucleotide triphosphates on this site stimulates both the
binding of the coenzyme and the reduction of certain substrate
triphosphates (26).
From the results described here we can not unambiguously
describe the mechanism by which Ado-Met fulfills its two
functions of allosteric effector and of methyl donor. At least two
hypotheses can be postulated. The first one involves the existence
of two different kinds of binding sites for Ado-Met on Dam
methylase. On a small monomeric enzyme of 32 kD, this would
be surprising, but not impossible. The alternative hypothesis
would be that there is a single binding site for Ado-Met. In the
absence of DNA, binding at this site induces an allosteric effect.
When the DNA is bound to the enzyme/Ado-Met complex further
conformational rearrangement takes place. This could involve
either the transformation of the allosteric site into a catalytic one
or the creation of a second site.
ACKNOWLEDGEMENTS
A.B. held a 'Contrat Formation Recherche' of the C.E.A. during
the course of this work. This work was partially supported by
a Twinning grant from the European Community
(ST27-0034-1F). A. Schulz, T. A. Trautner, G. Herve, B.
Labouesse, G.V. Fazakerley, F. Delia Seta, and D. Thiele are
acknowledged for their helpful discussions. We thank N. O.
Reich and N. Mashoon for personnal communications.
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
Geier, G. E. and Modrich, P. (1979) J. Biol. Chem. 254, 1408-1413
Ogden, G. B., Pratt, M. J. and Schaechter, M. (1988) Cell 54, 127-135
Messer, W. and Noyer-Weidner, M. (1988) Cell 54, 735-737
Barras, F. and Marinus, M. G. (1989) Trends in Genetics 5, 139-143
Herman, G. E. and Modrich, P. (1982) J. Biol. Chem. 257, 2605-2612
Bergerat, A., Kriebardis, T. and Guschlbauer, W. (1989) J. Biol. Chem.
264, 4064-4070
Jack, W.E., Rubin, R. A., Newman, A. and Modrich, P. (1980) in 'Gene
Amplification and Analysis 1: Restriction Endonucleases', Chirikjian, J. G.,
sEd., Elsevier-North Holland, p. 165-170
Hadi, S. M., Bickle, Th. A. and Yuan, R. (1975) J. Biol. Chem. 250,
4159-4164
Fritz, H. J., Belagaye, R., Brown, E., Fritz, R., Jones, R., Less, R. and
Khorana, H. G. (1978) Biochemistry 17, 1257-1267
Pulleyblank, D., Midialak, M., Laurent-Daisley, St. and Glick, R. (1983)
Mol. Biol. Rep. 9, 191-195
11. Wagner, J., Danzin, C. and Mamont, P. (1982) J. Chromatogr. 227,
349-368
12. Kriebardis, A. and Guschlbauer, W. (1987) FEBS-Letters 213, 297-300
13. Bradford, M. M. (1976) Anal. Biochem. 72, 248-252
14. Huet, J., Cotrelle.P., Cool, M., Vignais, M. L., Thiele, D., Marck, C ,
Buhler, J. M., Sentenac, A. and Fromageot, P. (1985) EMBO J. 14,
3539-3547
15. Fersht, A. (1984) Enzyme Structure and Mechanism, W. H. Freeman, San
Fransisco, p. 150—151
16. Bading, H. (1988) Nucleic Acids Res. 16, 5241 -5249
17. Fazakerley, G.V. (personal communication)
18. Wenzel, C., Moulard, M. and Guschlbauer, W. (unpublished work)
19. Som, S. and Friedman S. (1990) J. Biol. Chem. 265, 4278-4283
20. Wu, J. C. and Santi D. V. (1987) J. Biol. Chem. 262, 4778-4786
21. Friedman, S. (1986) Nucleic Acids Res.14, 4543-4556
22. Schulz, A. and Trautner, T. A. (personal communication)
23. Reich, N. O. and Mashhoon, N. (personal communication)
24. Steitz, T. A., Harrisson, R., Weber, I.T. and Leahy, M. (1982) in 'Mobility
and function in proteins and nucleic acids'. Pitman, London (Ciba Foundation
Symposium 93), p. 25-46
25. Gersonde, K., Sick, H., Overkamp, M., Smith, K.M. and Parish, D. W.
(1986) Eur. J. Biochem. 157, 393-404
26. Eriksson, S. and SjUberg, B. M. (1989) in Allosteric Enzymes, Herve\ G.,
Ed., CRC Press, Boca Raton, FL, p. 189-215