|
|
|
|||||||
|
|||||||||
(Received for publication, September 22, 1994; and in revised form, December 18, 1994) From the
The restriction endonuclease EcoRI binds and cleaves
DNA containing GAATTC sequences with high specificity. According to the
crystal structure, most of the specific contacts of the enzyme to the
DNA are formed by the extended chain region and the first turn of
One century ago the concept of complementarity between an enzyme
and its substrate was introduced by Emil Fischer (1894) (
Figure 1:
A, EcoRI-DNA co-crystal
structure (Kim et al., 1990; Brookhaven Data Bank entry 1R1E).
The region of the extended chain-
The majority of the specific contacts between EcoRI
and DNA are formed by amino acids within a short region of the protein,
namely the extended chain-
Figure 2:
Cleavage of
oligoRI by EcoRI in the presence of 100 µM peptide (
Figure 3:
A, cleavage of
Figure 4:
Rates of
Figure 5:
A,
cleavage of pRVIS1 plasmid with XbaI in the absence of peptide (- peptide) as well as in the presence of 500 µM peptide (+ peptide). The supercoiled plasmid (sc) is cleaved via an open circle intermediate (oc)
to give the linear DNA (lin). 5 µg of plasmid were
incubated with 5 units of XbaI in EcoRI cleavage
buffer at 37 °C. After the times given (`, min), aliquots
were withdrawn and analyzed. B, cleavage of pRIF309+
plasmid with XbaI in the absence of peptide (- peptide) as well as in the presence of 500 µM peptide (+ peptide). The supercoiled plasmid is cleaved to
produce two fragments 4153 and 915 bp in length. 3 µg of plasmid
were incubated with 5 units of XbaI in EcoRI cleavage
buffer at 37 °C. After the times given (`, min), aliquots
were withdrawn and analyzed.
Figure 6:
Gel electrophoretic mobility-shift assay
of EcoRI in the absence of peptide (- peptide)
as well as in the presence of 250 and 500 µM peptide. DNA
was incubated with 0, 15, 25, 50, 75, or 150 nMEcoRI (left to right) at 21 °C in binding
buffer.
Figure 7:
Methylation of
The concentration of both EcoRI and the
macromolecular DNA substrates used in the cleavage experiments
described above was low. Moreover, a high excess of nonspecific
competitor sites is present in the reaction mixture when macromolecular
DNAs are employed as substrates (Langowski et al., 1980).
Under these conditions substrate binding is the rate-limiting step of
the reaction (Langowski et al., 1981). Recently, we have shown
that EcoRI, when diffusing along the DNA, does not miss a
recognition site under conditions similar to those employed here
(Jeltsch et al., 1994). This implies that the rate of
formation of the enzyme-EcoRI site complex governs the
observed reaction rate. Therefore, the overall rate can be described by
a second order rate equation, v =
kc
where v The relative rates
determined at various peptide concentrations, therefore, can be used
directly to calculate the DNA binding constant of the peptide. This
analysis was carried out with the computer program TITRAT, which
calculates an association constant with a least square fit method using
a multistep predictor/corrector module (VA05A) (Powell, 1965). The
binding constant turned out to be 3 The cleavage of
oligoRI was carried out at relatively high substrate concentrations
(0.5 µM). Fluorescence-stopped flow studies with oligoRI
have shown that substrate binding of EcoRI occurs in a
pre-equilibrium kinetically separable from cleavage under these
conditions (Alves et al., 1989b). Therefore, the
Michaelis-Menten model is applicable to analyzing the kinetics of
oligoRI cleavage, v = c The effect of the peptide on the rate
of DNA cleavage by EcoRI under these conditions is to reduce
the concentration of free oligoRI leading to a decrease in rate. The
ratio of rates measured in the absence and in the presence of 100
µM peptide (v
where c The DNA binding reaction of EcoRI in the presence
of peptide is governed by two coupled equilibria: EcoRI
binding to the DNA and peptide binding to the DNA. For a quantitative
analysis of the gel electrophoretic mobility-shift assays (Fig. 6), the lanes with the two highest EcoRI
concentrations were analyzed to estimate the DNA binding affinity of EcoRI and of the peptide. The fractions of DNA bound by EcoRI in the lanes without peptide yield a binding constant of EcoRI to the specific site of 1.5
Specific interactions of proteins with DNA have been
investigated in great detail in several cases. Often specific contacts
between the proteins and the DNA are formed by characteristic
structural elements (for reviews see Steitz(1990) and Harrison(1991)).
Frequently Often the DNA-binding regions are stable subdomains that
specifically interact with the DNA as demonstrated for basic region
leucine zipper, helix-turn-helix, basic region helix-loop-helix, and
zinc finger proteins. Indeed many of the available structures of
DNA-binding proteins were determined only with the DNA-binding domain
or subdomain of the protein. An independent folding of the DNA binding
module, however, could not be expected for EcoRI, because the
extended chain- Interestingly, the discrimination of the peptide
between GAATTC and CAATTG is less stringent than the discrimination of EcoRI between these sequences. This observation is in
accordance with the recognition scheme, which is based on the structure
of the specific EcoRI-DNA co-crystal (Rosenberg, 1991),
because, in contrast to the AT base pairs, the two symmetry-related GC
base pairs are contacted mainly by amino acids outside of the extended
chain- Specific
binding of short peptides to nucleic acids is not a novel phenomenon.
For the basic region leucine zipper protein GCN4, it was shown that a
small peptide comprising 20 residues specifically interacts with DNA
(Talanian et al., 1992). Specific DNA interaction of minor
groove-binding peptides is observed with short peptides containing RGR
repeats, which resemble minor groove-binding drugs like netropsin or
distamycin (Geierstanger et al., 1994). Moreover, a peptide 17
amino acids in length containing the Arg-rich region of the HIV Rev
protein is able to bind to the Rev response element in RNA (Tan et
al., 1993). Other examples of RNA-binding protein motifs were
reviewed recently (Mattaj, 1993). The binding of the extended
chain- The binding affinity of the peptide to the GAATTC sequence
was determined to be 3 The binding affinity of the peptide to CAATTG sequences
has been estimated to be 10 Most of the specific contacts of EcoRI to the bases
of its recognition sequence (GAATTC) are formed by a short continuous
peptide sequence, the extended chain-
Volume 270,
Number 10,
Issue of March 10, 1995 pp. 5122-5129
©1995 by The American Society for Biochemistry and Molecular Biology, Inc.
4 Region of the
Restriction Endonuclease EcoRI Specifically Binds to the EcoRI Recognition Site (*)
ABSTRACT
INTRODUCTION
MATERIALS AND METHODS
RESULTS
DISCUSSION
CONCLUSIONS
FOOTNOTES
ACKNOWLEDGEMENTS
REFERENCES
-helix 4 (amino acids 137-145). Here, we demonstrate
that a dodecapeptide (WDGMAAGNAIER), which is identical in the
underlined parts of its sequence to EcoRI amino acids
137-145, specifically binds to GAATTC sequences. The peptide
inhibits DNA cleavage by EcoRI but not by BamHI, BclI, EcoRV, HindIII, PacI, and XbaI. DNA cleavage by XbaI is slowed down at sites
that partially overlap with EcoRI sites. The peptide inhibits
cleavage of GAATTC sites by ApoI, which recognizes the
sequence RAATTY. It interferes with DNA methylation by the EcoRI methyltransferase but not by the BamHI
methyltransferase. It competes with EcoRI for DNA binding.
Based on these results, the DNA binding constant of the peptide to
GAATTC sequences was calculated to be 3 
10
M
. DNA binding is not
temperature-dependent, suggesting that binding of the peptide is
entropy-driven. As the peptide does not show any nonspecific binding to
DNA, its DNA binding specificity is similar to that of EcoRI,
in spite of the fact that the affinity is much smaller. These results
suggest that contacts to the phosphate groups in EcoRI mainly
provide binding affinity, whereas the specificity of EcoRI is
based to a large extent on sequence-specific base contacts.
)to
explain the specificity of enzymes. This concept has proven to be one
of the most successful concepts in enzymology and has been demonstrated
to be applicable in numerous cases including very specific enzymes, for
example type II restriction endonucleases. These enzymes (for reviews
see Heitman (1993) and Roberts and Halford(1993)) recognize palindromic
sequences 4-8 bp (
)in length and cleave the DNA within
these sequences. As has been shown, for example, for EcoRI
(recognition site, GAATTC), sequences differing in only 1 base pair
from the canonical sequence (``star'' sites) are cleaved at
least 3 orders of magnitude more slowly (Lesser et al., 1990;
Thielking et al., 1990), and sites differing in more than 1
base pair are not cleaved at all (Gardner et al., 1982;
Rosenberg and Greene, 1982). Similarly, binding of star sites is
impaired by at least 2 orders of magnitude when compared with binding
of GAATTC sequences; other sites are bound at least 4 orders of
magnitude more weakly than GAATTC (Lesser et al., 1990;
Thielking et al., 1990). The structural basis of this high
specificity is explained by the x-ray structure analysis of a specific EcoRI-DNA co-crystal (see Fig. 1) (McClarin et
al., 1986; Kim et al., 1990). It demonstrated that EcoRI binds as a symmetrical dimer to the palindromic
recognition site and identified in the protein-DNA interface many
specific contacts between the protein and the DNA (Rosenberg, 1991; Kim et al., 1993). They are formed between the bases of the GAATTC
sequence and EcoRI (direct readout) as well as between the
phosphate groups of the DNA backbone and the protein (indirect
readout). All specific contacts to the bases of the recognition
sequence are compiled in Fig. 1B, namely nine hydrogen
bonds to the bases of the DNA (two of which are mediated by a water
molecule) and five hydrophobic contacts. Additionally, at least nine
phosphate contacts are observed. Interestingly, direct readout is
almost exclusively due to a short region of EcoRI, the
extended chain motif (Met
-Ala
), which is
deeply buried in the major groove of the DNA, and the amino-terminal
part of -helix 4 (Ile
-Arg
). This
extended chain-4 region (Met-Arg)
forms all direct (i.e. not water-mediated) hydrogen bonds and
three of five hydrophobic contacts. Taken together, on the basis of the
x-ray structure analysis and many biochemical studies (Brennan et
al., 1986; Fliess et al., 1986; McLaughlin et
al., 1987; Needels et al., 1989; King et al.,
1989; Alves et al., 1989a; Heitman and Model, 1990a, 1990b;
Osuna et al., 1990; Oelgeschläger et
al., 1990; Jeltsch et al. 1993a), it appears as if the
specificity of EcoRI is based on an extensive complementarity
of the extended chain-4 region and the major groove of the GAATTC
sequence. Here we have directly tested this assumption by investigating
the DNA binding properties of a short peptide with an amino acid
sequence identical to the extended chain-4 region.
4 peptide is highlighted (thick line). B, schematic drawing of the specific
base contacts observed in the EcoRI-DNA co-crystal structure
(Rosenberg, 1991; Kim et al., 1993). Contacts of one subunit
to the GAATTC sequence are indicated, most of which are formed by the
extended chain-4 region (Met-Arg).
Peptide Synthesis
Peptide synthesis was
performed on a MilliGen 9050 peptide synthesizer by an automated method
as reported earlier (Zhang et al., 1990). An HPLC analysis of
the crude product mixture showed only one major peak. A few minor peaks
contained less than 5% intensity. The peptide was purified by HPLC
(Nucleosil C4, 5 µm) to homogeneity. The identity of the peptide
was confirmed by ion spray mass spectroscopy. The concentration of the
peptide was determined using a molar extinction coefficient
= 5540 M cm, which is a mean value for tryptophan
absorbance in polypeptides (Mach et al., 1992).Oligonucleotide Synthesis
The self-complementary
oligodeoxynucleotide oligoRI (TATAGAATTCTAT) was synthesized and
purified as described (Alves et al., 1989b).DNA Preparation
Plasmid DNA was prepared using
Midi- or Maxi-preparation kits (QIAGEN Inc.) according to the
instructions of the supplier. Supercoiled pUC8 was prepared by CsCl
density gradient ultracentrifugation (Sambrook et al., 1989).
-DNA was supplied by Boehringer Mannheim. A radioactively labeled
62-bp shift-substrate with one EcoRI site was generated by
primer extension in a polymerase chain reaction containing
[-
P]dATP.Restriction and Modification
Enzymes
EcoRI (recognition site, GAATTC) and EcoRV (GATATC) were homogeneous preparations from
overproducing Escherichia coli cells (Geiger et al.,
1989; Fliess et al., 1988). BamHI (GGATCC) and MunI (CAATTG) were supplied by the U. S. Biochemical Corp.
(Cleveland, OH) and Fermentas (Vilnius, Lithuania), respectively. XbaI (TCTAGA), BclI (TGATCA), and HindIII
(AAGCTT) were obtained from Angewandte Gentechnologie Systeme
(Heidelberg, Germany). ApoI (RAATTY), PacI
(TTAATTAA), EcoRI methyltransferase, and BamHI
methyltransferase were from New England Biolabs (Beverly, MA).Oligodeoxynucleotide Cleavage Experiments
OligoRI
cleavage by EcoRI was followed with a continuous
spectrophotometric assay that is based on the hyperchromicity of DNA
(Waters and Connolly, 1992). Cleavage experiments were carried out in EcoRI buffer (20 mM Tris
HCl, pH 7.5, 10 mM MgCl, 50 mM NaCl) in microcuvettes (100
µl) in a Hitachi U3210 photometer at the temperatures indicated
(4-25 °C). Reaction progress curves were stored and analyzed
by linear regression of the initial part of the curve. OligoRI
concentrations of 0.5 µM and EcoRI concentrations
of 14.8 nM were used. DNA cleavage rates (A
/min) were determined from the linear part
of the reaction progress curves in the absence or in the presence of
100 µM extended chain-4 peptide.-DNA Cleavage Experiments-DNA cleavage
experiments in the absence and presence of the extended chain-4
peptide were carried out with EcoRI, EcoRV, BamHI, MunI, HindIII, and BclI in
order to determine whether the enzymes are inhibited by the peptide.
All reactions were performed in the same buffer (EcoRI buffer:
20 mM TrisHCl, pH 7.5, 10 mM MgCl,
50 mM NaCl) to ensure comparable binding conditions for the
peptide to the DNA. If not otherwise stated, reactions were carried out
at 37 °C. Usually 2 µg of DNA were cleaved with 0.5-3
units of the restriction enzyme. Peptide concentrations were varied
between 0 and 500 µM as indicated. After appropriate times
aliquots containing 0.25 µg of DNA were withdrawn, stopped by
addition of gel loading buffer (0.1 M EDTA, 25% v/v Ficoll,
0.2% w/v bromphenol blue, 0.2% w/v xylene cyanol), and analyzed on 0.8%
w/v agarose gels. For the analysis of cleavage rates in the absence and
in the presence of peptide, the intermediate patterns of cleavage
products of the reactions were compared.Plasmid-DNA Cleavage Experiments
The plasmid
pRIF309+ (5068 bp) was cleaved with EcoRI (three sites,
fragments 2547 bp, 1598 bp, and 923 bp), XbaI (two sites,
fragments 4153 bp and 915 bp), and PacI (one site). pRVIS1
(4607 bp) was cleaved with XbaI (one site). Plasmid pUC8 (2665
bp) was cleaved with EcoRI (one site) and ApoI (one
site). Reactions were carried out in the absence and presence of the
extended chain-4 peptide and analyzed as described for the
-DNA cleavage reactions.Mobility-shift Experiments
Gel-shift experiments
with EcoRI and a 62-bp DNA fragment containing one EcoRI site were carried out in the absence as well as in the
presence of 250 and 500 µM peptide to find out if DNA
binding by EcoRI is affected by the addition of extended
chain-4 peptide. DNA (100 nM) was mixed with EcoRI in binding buffer (50 mM TrisHCl, pH 7.5,
50 mM NaCl, 10 mM 2-mercaptoethanol, 2 mM spermine, 0.1 mg/ml bovine serum albumin, 2 mM EDTA) and
incubated for 30 min at 21 °C. Subsequently, 3 µl of gel
loading buffer (50% glycerol, 50 mM TrisHCl, pH 7.5, 50
mM NaCl, 10 mM 2-mercaptoethanol , 2 mM spermine, 0.1 mg/ml bovine serum albumin, 2 mM EDTA, 0.2%
w/v bromphenol blue, 0.2% w/v xylene cyanol, 0.2% w/v azorubin) were
added to a 5-µl aliquot of the binding mixes, and the samples were
subjected to gel electrophoresis on 6% w/v polyacrylamide gels at 150 V
in 0.5 TTE (90 mM Tris, 29 mM taurine, 1.25 mM EDTA). Gels were dried, and the radioactivity of the spots was
analyzed.DNA Methylation Kinetics
DNA methylation kinetics
in the presence and absence of the extended chain-4 peptide were
carried out to determine if the EcoRI methylase, like the EcoRI restriction endonuclease, is inhibited by the peptide.
-DNA (2 µg) was mixed with 5 units of EcoRI methylase
in 50 mM TrisHCl, pH 8.0, 100 mM NaCl, 10
mM EDTA, 8 µMS-adenosylmethionine in
the absence and presence of 500 µM extended chain-4
peptide in 20 µl of reaction volume. After appropriate times,
aliquots were withdrawn and added to EcoRI buffer containing EcoRI. In this second reaction non-methylated DNA was
digested. In order to ensure digestion even in the presence of peptide,
a high concentration (100 units) of EcoRI was used, and the
samples were incubated for 45 min at 37 °C. Methylation kinetics
with BamHI methylase were performed similarly as described for
the EcoRI methylase, except that the analytical digest was
carried out with 20 units of BamHI in BamHI cleavage
buffer. Control experiments showed that, under these conditions,
non-methylated DNA is completely cleaved, regardless of whether the
peptide is present in the reaction mixture or not.Biophysical Methods
Sedimentation equilibrium runs
(60,000 rpm; up to 72 h) were carried out in a Spinco model E
analytical ultracentrifuge (Beckman) equipped with an electronic spin
control, a high-intensity UV illumination system, a photoelectric
scanner, and an electronic multiplexer. Concentration profiles were
recorded at 280 or 294 nm (depending on the concentration of the
peptide) and stored in a computer. Molecular weights were determined
from a log(c(r)) versusr plot. Concentrations of the peptide were varied in the runs
between 18 and 90 µM. CD spectra of 0.1 mg/ml peptide in EcoRI buffer were recorded at ambient temperature in 0.01-cm
cuvettes in a Jobin-Yvon Dichrograph R. J. Mark III calibrated with
(+)-camphor-10-sulfonic acid, D-pantolactone, and
epiandrosterone. Spectra were recorded in a computer and analyzed
numerically in terms of -helix and 
-sheet content (Chen et al., 1972). CD melting curves were recorded at 200 nm from
15 to 80 °C with a temperature rise of 20 °C/h.
4 region
(Met-Arg) (Fig. 1). Here we have
studied the DNA binding activity of a short dodecameric peptide
(HN-WDGMAAGNAIER-COOH), which is identical in sequence to
the amino acids Asp
-Arg in EcoRI
except for the amino-terminal Trp that was added to allow for
spectroscopic determination of the concentration of the peptide and for
a Leu
Gly replacement, which was made in order to
increase the solubility of the peptide.Biophysical Characterization of the Peptide in
Solution
The peptide turned out to be soluble in water up to
concentrations of 2 mM. It tends, however, to precipitate
after repeated freezing/thawing cycles. The analysis of sedimentation
equilibrium runs in 20 mM TrisHCl, pH 7.5, 50 mM KCl, 10 mM MgCl resulted in an apparent
molecular weight of 1180 ± 70 g mol comparable with the theoretical value of 1290 g
mol. Therefore, the peptide is a monomer in
solution. CD spectra (250-190 nm) of the peptide in 20 mM TrisHCl, pH 7.5, 50 mM NaCl, 10 mM MgCl demonstrate that the peptide does not adopt a
regular structure (-helix or -strand) in solution, a
conclusion that is supported by the absence of a melting transition
between 15 and 80 °C.Experimental Strategy
To find out whether the
peptide can specifically interact with the recognition sequence of EcoRI (GAATTC), we measured the inhibition of the activity of
several restriction and modification enzymes by the peptide using
various DNA substrates. These inhibition experiments resemble classical
footprinting analyses in which the DNA is protected against the attack
of a nonspecific nuclease- or DNA-degrading chemical by ligands bound
to the DNA. As such, they are indirect measurements that allow one to
detect unequivocally even weak binding of the peptide to the DNA. Other
standard techniques to detect DNA protein interactions, viz. gel-shift experiments, nitrocellulose filter binding, fluorescence
(including fluorescence-detected stopped flow and fluorescence-detected
temperature jump), circular dichroism, and NMR spectroscopy, failed to
demonstrate DNA binding of the peptide, presumably because the complex
is not sufficiently stable and/or because complex formation does not
result in a spectroscopic signal change.Inhibition of DNA Cleavage by EcoRI
First, we
measured the influence of the extended chain-4 peptide on the
cleavage of a 13-mer oligodeoxynucleotide (oligoRI) by EcoRI.
As an example, the reaction progress curves in the absence and presence
of 100 µM peptide are shown in Fig. 2. The initial
rate of DNA cleavage by EcoRI is reduced by a factor of 1.4
± 0.1 in the presence of 100 µM peptide. We then
used several macromolecular substrates for the cleavage inhibition
experiments. It turned out that EcoRI cleavage of -DNA,
pUC8-DNA, and pRIF309+ plasmid DNA is strongly inhibited by the
addition of peptide at concentrations higher than 50 µM.
Examples of the time course of -DNA and pUC8-DNA cleavage by EcoRI in the absence of peptide as well as in the presence of
500 µM peptide are shown in Fig. 3. The relative
rates of -DNA and pUC8-DNA cleavage by EcoRI at various
peptide concentrations are shown in Fig. 4. These results
demonstrate that the peptideinterferes with DNA cleavage by EcoRI, either by binding to the DNA or to the enzyme.
, right ordinate) as well as in
the absence of peptide (-, left ordinate) in EcoRI cleavage buffer at 25 °C. In each case 0.5
µM oligoRI was cleaved with 14.8 nMEcoRI. The initial rates of cleavage (straight
line) differ by a factor of 1.4.
-DNA in the absence (- peptide) of peptide as well as in the presence of 500
µM peptide (+ peptide) in EcoRI
cleavage buffer at 37 °C. 2 µg of -DNA were incubated with
0.5 nMEcoRI in 70 µl of reaction volume (c
sites = 5.4
nM). After the times given (`, min),
aliquots were withdrawn and analyzed electrophoretically. -DNA
(48,502 bp) is cleaved by EcoRI into six fragments (21,225,
7,421, 5,804, 5,643, 4,878, and 3,531 bp). B, cleavage of
pUC8-DNA in the absence (- peptide) of peptide as well
as in the presence of 500 µM peptide (+ peptide) in EcoRI cleavage buffer at 37 °C. 1 µg
of pUC8-DNA was incubated with 0.1 nMEcoRI in 100
µl of reaction volume (c sites = 5.7 nM). After the times given (`, min), aliquots were withdrawn and analyzed
electrophoretically. The supercoiled pUC8-DNA (sc) is cleaved
via an open circle intermediate (oc) to give the linear form (lin).
-DNA and pUC8-DNA cleavage
by EcoRI at various concentrations of peptide. All rates are
given relative to the rates of DNA cleavage of identical samples
measured in the absence of peptide. Rates were determined in EcoRI buffer at 21 °C. Values given are accurate within
±20%. The line is a theoretical curve corresponding to
the best fit of the data (K
(K
) = 3
10M). The dashed lines are simulated curves for K =
1.5 10M (upper) and K = 6
10M (lower),
respectively, and are included to demonstrate that 2-fold variations of
the K result in a significantly worse
fit.
Inhibition of DNA Cleavage by Other Restriction
Enzymes
All results obtained with inhibition studies of
different enzymes by the peptide are compiled in Table 1. ApoI (RAATTY), like EcoRI, cleaves GAATTC sequences.
The ApoI cleavage of plasmid pUC8 that contains one GAATTC
sequence but no additional ApoI sites was inhibited
approximately 20-fold in the presence of 500 µM peptide.
This value is nearly identical to that obtained for the inhibition of
the EcoRI-catalyzed cleavage of pUC8 (17-fold). XbaI
(TCTAGA)-catalyzed cleavage of plasmid pRVIS1 was not inhibited by 500
µM peptide in the reaction buffer. In contrast, cleavage
of pRIF309+ by XbaI was inhibited by addition of the
peptide (Fig. 5). This is an important result, because in
pRIF309+ both XbaI sites are part of a polylinker and
partially overlap with EcoRI sites (TCTAGAATTC).
These findings taken together demonstrate that the inhibition of EcoRI by the peptide is not due to an interaction of the
peptide with EcoRI but rather with the DNA.
Specificity of the Peptide-DNA Interaction
To
determine the specificity of the peptide-DNA interaction we have
measured whether the peptide also inhibits several other restriction
endonucleases (Table 1). -DNA cleavage by EcoRV
(GATATC), BamHI (GGATCC), HindIII (AAGCTT), or BclI (TGATCA) was not influenced by the addition of 500
µM peptide to the reaction mixtures. Similarly, PacI (TTAATTAA) cleavage of pRIF309+ and XbaI
(TCTAGA) cleavage of pRVIS1 were not inhibited by the peptide.
Consequently, the peptide does not bind to the sequences GATATC,
GGATTC, AAGCTT, TAATTA, TGATCA, and TCTAGA, some of which are closely
related to GAATTC. In order to test the selectivity with a substrate in
which the outer base pairs of the recognition sequence were changed, we
measured the DNA cleavage rate of MunI (CAATTG) in the
presence of the peptide. In contrast to all other enzymes tested,
except for EcoRI and ApoI, MunI turned out
to be inhibited by the peptide, albeit more weakly than EcoRI,
because 500 µM peptide only resulted in a 1.5-fold
inhibition of the cleavage rate. To estimate the binding of the peptide
to EcoRI star sequences, we carried out inhibition experiments
of EcoRI star cleavage with -DNA as substrate. Star
cleavage of EcoRI was induced either by using a
Mn
-containing cleavage buffer (20 mM TrisHCl, pH 7.5, 50 mM NaCl, 1 mM MnCl) (Hsu and Berg, 1978) or by using very high
enzyme concentration in cleavage experiments (up to 1 µM)
under normal buffer conditions. Under either condition, star activity
of EcoRI was inhibited to the same extent as the canonical
cleavage activity (data not shown), demonstrating that the peptide
discriminates between GAATTC and star sequences to a similar degree as EcoRI.Reduction of DNA Binding by EcoRI
As an
independent test of the interaction between the peptide and GAATTC, the
influence of the peptide on DNA binding by EcoRI was
investigated. We employed a gel retardation assay, which is well suited
to compare relative binding affinities. DNA binding experiments with EcoRI were carried out in the absence of peptide as well as in
the presence of 250 and of 500 µM peptide, respectively,
in the binding mixtures. Fig. 6shows an example of the results.
Clearly, the peptide competes with EcoRI for binding to the
specific DNA sequence, because, at the highest EcoRI
concentration used, approximately 85% of the DNA are shifted in the
absence of peptide, but only 50% are shifted in the presence of 500
µM peptide.
Inhibition of DNA Methylation by the EcoRI
Methylase
DNA methylation activity of the EcoRI
methyltransferase was measured in the absence of the extended
chain-4 peptide as well as in the presence of 500 µM peptide. As shown in Fig. 7, the activity of the methylase
is reduced approximately 4-fold in the presence of peptide, as
indicated by the smaller amount of methylated (i.e. uncleavable) DNA when the methylation was carried out in the
presence as compared with the absence of peptide. The activity of the BamHI methyltransferase, in contrast, was not affected by the
presence of 500 µM peptide in the reaction mixture.
-DNA by the EcoRI methyltransferase. 2 µg of -DNA were incubated
with 5 units of methylase in reaction buffer at 37 °C in the
absence (- peptide) as well as in the presence (+ peptide) of 500 µM peptide. After the
times indicated (`, min), aliquots were withdrawn and analyzed
by digestion with EcoRI.
Estimation of the Peptide-DNA Binding
Constant
Taken together, the data presented so far can only be
explained if the peptide specifically interacts with the EcoRI
recognition sequence (GAATTC), albeit with a low but clearly detectable
affinity. The dependence of the inhibition measured for DNA cleavage
and binding by EcoRI on the peptide concentration can be used
to calculate the equilibrium constant for the binding of the peptide to
the DNA.
c
, where c is the concentration of DNA and c is the
concentration of EcoRI. Assuming that EcoRI cannot
cleave the DNA in the DNA-peptide complex, under identical conditions
the observed reaction rate is proportional to the fraction of the DNA
not complexed with peptide,
is the initial
cleavage velocity in the presence of peptide, v is
the initial cleavage velocity in the absence of peptide,
c
is the concentration of
free DNA in the presence of peptide, and
c
is the total
concentration of DNA in the presence of peptide. 10M (Fig. 4). With the same
analysis, the equilibrium binding constant for the binding of the
peptide to the sequence CAATTG, which was derived from inhibition of MunI-catalyzed DNA cleavage, was estimated to be 1
10
M.k
c/K
+ c, where k is the turnover number.
=
v/v = 1/1.4) (Fig. 2) together with K = 80 nM and k = 23 s (Jeltsch et al., 1993b) can be used to deter-mine the concentration of
free oligoRI in the presence of peptide
(c
),
is the concentration of
oligoRI in the absence of peptide, which is equal to the total
concentration of oligoRI. With
c, the concentration of
oligoRI-peptide complexes and finally K
for the
peptide-DNA complex can be calculated. With this procedure a K of 3 10M was obtained in agreement with the
result of the analysis of the inhibition of -DNA and pUC8-DNA
cleavage. 10
M under the conditions of the
experiments. This value, in combination with the concentration of the EcoRI-DNA complexes in the presence of peptide, can be used to
calculate the concentration of the free DNA in these mixtures. Then the
concentration of the peptide-DNA complexes and the equilibrium constant
for the binding of the peptide to the DNA can be estimated to be
2-4 10M, which
is similar to the values derived in the other analyses.Temperature Dependence of the Peptide-DNA
Interaction
Cleavage of oligoRI by EcoRI in the
presence of 100 µM peptide at both 4 and 25 °C
resulted in a reduction of the cleavage rate constant by the same
factor of 1.4. Similarly, EcoRI cleavage of pUC8-DNA and
-DNA was equally inhibited by the peptide at 4, 21, and 37 °C
(data not shown).MunI Peptide DNA Interaction
Recently, it has been
suggested on the basis of sequence alignments that the restriction
enzyme MunI has a structure similar to that of EcoRI
and, in particular, that a S
AGRGNAHER region in MunI has a similar function as the extended chain-4
region in EcoRI, namely to recognize the DNA (Siksnys et
al., 1994). We have therefore also prepared a MunI
peptide (WDPSAGRGNAHER), which corresponds to the amino acid sequence
of the putative extended chain region in MunI. This peptide,
however, inhibited neither EcoRI nor MunI (data not
shown). Although these results do not support the suggestion of
structural and functional similarities between the two enzymes, they do
not disprove them either. However, these results clearly demonstrate
that a peptide of similar length and sequence to the extended
chain-4 peptide binds neither to GAATTC nor to CAATTG sequences.
It is noteworthy in this context that an EcoRI mutant in which
the extended chain-4 region was replaced by the corresponding
region of MunI by site-directed mutagenesis of the EcoRI gene was catalytically inactive (Fritz, 1994).
-helices are positioned in the major groove of the DNA, e.g. by helix-turn-helix, basic region helix-loop-helix, basic
region leucine zipper, and zinc finger proteins, to provide a
structural framework for the recognition interactions (for a recent
review see Wolberger(1993)) but -sheets, as in the MetJ- and
Arc-repressors, are also employed (for a recent review see Rauman et al.(1994)). In the EcoRI-DNA complex, a short
segment of the protein comprising amino acids
Met-Arg forms nearly all specific contacts
to the bases of the recognition sequence. This segment has an extended
conformation and is deeply buried in the major groove of the DNA. 4 motif is held in place by several interactions to
other parts of the protein. Although the extended chain-4 region
of EcoRI, therefore, cannot be considered to be a subdomain,
here we demonstrate by several lines of evidence that a dodecameric
oligopeptide that contains this sequence specifically binds to GAATTC
sequences. These lines of evidence are as follows. (i) The peptide
inhibits the EcoRI-catalyzed DNA cleavage of several different
substrates (13-mer oligodeoxynucleotide, pUC8-DNA, pRIF309+
plasmid DNA, and -DNA) in a concentration-dependent manner. (ii)
The relative cleavage rates of GAATTC sites by EcoRI, which
recognizes GAATTC sequences and ApoI (recognition sequence,
RAATTY), are equally reduced by the peptide. (iii) DNA cleavage of BamHI (GGATCC), BclI (TGATCA), EcoRV
(GATATC), HindIII (AAGCTT), PacI (TTAATTAA), and XbaI (TCTAGA) is not affected by the peptide. (iv) XbaI cleavage at sites that partially overlap with EcoRI sites (TCTAGAATTC) is inhibited by the
peptide. (v) DNA methylation by the EcoRI methyltransferase
but not by the BamHI methyltransferase is slowed down by the
peptide. (vi) The peptide competes with specific DNA binding by EcoRI.4 region rather than by amino acids within this region (Fig. 1). The close contact of Met and Ala
to the GC base pair, on the other hand, appears sufficient to
discriminate a GC base pair from a TA base pair that contains a methyl
group in the major groove of the DNA at this position, because TTAATTAA
cleavage by PacI was not inhibited by the peptide.4 peptide to DNA demonstrated in this work differs from all
of these examples, because this peptide does not contain a net positive
charge or positive charge clusters, which could support nonspecific
binding to DNA via electrostatic contacts to the phosphate backbone.
Consequently, in the EcoRI-DNA co-crystal structure, the
extended chain-4 region is not involved in phosphate contacts to
the DNA. 10M, which corresponds to an
interaction energy of, 
G =
-RT ln K = -25.3 kJ
mol (-6.05 kcal mol). The
binding constant of the peptide to GAATTC turned out to be
temperature-independent between 4 and 37 °C, within the limits of
error of our experiments. This result shows that H of the peptide-DNA association is small and, hence, that the
reaction is mainly entropy-driven. This is surprising at first, because
the peptide is most likely disordered in solution but presumably well
ordered in the complex with the DNA. One has to expect, therefore, the
existence of the unfavorable entropy term
S
upon complex formation. One
might speculate that this term is overcompensated by the favorable term
S
that arises, because upon complex
formation 1450 Å of solvent-accessible surface are
buried. Because 1 Å contributes roughly -100 J
mol (Chothia, 1974) to the interaction energy,
G can be estimated to be around
-145 kJ mol (-34.7 kcal
mol). The peptide contains 43 rotatable bonds that
generate rotational isomers (24 freely rotatable backbone bonds and 19
aliphatic C-C, C-N, or C-S bonds of the side chains).
Assuming that in solution only three rotational states are populated (i.e. 
= 60, 180, and 300°) and that rotation
is completely frozen in the complex, the contribution of the reduced
conformational flexibility to the entropy change of complex formation
can be estimated using S = k ln W,
where W denotes the number of possible conformations,
S = R ln W
/W
= R ln 1/(3
) = -390 J
K. Then, G can
be estimated to be 116 kJ mol (27.8 kcal
mol). Although G and G are only crude
estimates (in S, for example, an
altered flexibility of the DNA is not taken into account), the sum of
both terms is close to the G observed. This
estimation shows that the release of ordered water molecules could be
the thermodynamic driving force of the peptide-DNA association (Ha et al., 1989). What then is the function of the specific
hydrogen bonds? Before complex formation all hydrogen bond donors and
acceptors of the peptide and the DNA interact with water molecules. If
both surfaces match perfectly, all hydrogen bond donors and acceptors
are saturated after complex formation, too, resulting in a very small
net enthalpy change. If, however, the surfaces are not chemically
complementary to each other, some hydrogen bond donors or acceptors
would remain free but without access to water in the DNA-protein
interface yielding a large and positive H. This would
prevent association to such (i.e. nonspecific) sites, because
the overall G of the complex formation would become
positive.M.
A nonspecific binding affinity to other sequences was not detectable.
Given the sensitivity of the experiments, the binding affinity to
nonspecific sites can be estimated to be below 10M. Because this peptide contains
neither a positive net charge nor positive charge clusters, there is no
structural basis for a nonspecific binding of the peptide to DNA. The
specificity of the peptide in the discrimination of GAATTC and
nonspecific sequences, hence, is in the order of
10-10. This value is similar to the
discrimination factor of EcoRI in binding GAATTC and
nonspecific sequences (e.g.K(GAATTC)/K(CTTAAG)
= 7 10) (Lesser et al., 1990)
measured, however, in the absence of Mg. This
comparison demonstrates that the extended chain-4 region in EcoRI provides the major contribution to the binding
specificity of the enzyme. This conclusion is further supported by the
finding that cleavage at EcoRI star sites is similarly
inhibited by the peptide as cleavage at canonical sites.
4 motif. As demonstrated by
the crystal structure (Kim et al., 1990), this amino acid
motif is largely complementary to the major groove of the GAATTC
sequence, which enables it to form a variety of specific contacts to
the DNA; 8 amino acid residues (Met-Arg)
are involved in 10 specific interactions with the bases of the
recognition sequence. Here, we have demonstrated that a dodecapeptide
containing the sequence of the extended chain-4 motif specifically
binds to GAATTC sequences. Therefore, binding specificity of EcoRI is based mainly on the specific contacts between a small
amino acid sequence motif and the bases of the DNA, whereas binding affinity is provided by contacts between amino acid residues
dispersed over the entire DNA binding site of the protein and the
phosphate groups of the DNA. Our data suggest that, at least for the EcoRI restriction endonuclease, direct readout is more
important to ensure binding specificity than indirect readout. It must
be kept in mind, however, that cleavage specificity of restriction
enzymes is only in part due to binding specificity. It might well be
that in the transition state contacts to the phosphate groups play an
important role and determine whether a sequence is cleaved or not
(Koziolkiewicz and Stec, 1992; Jeltsch et al., 1993c).
)), association constant; G, Gibbs free energy of association; H, enthalpy of association; HPLC, high
pressure liquid chromatography; oligoRI, d(TATAGAATTCTAT).
We thank Dr. D. E. Wemmer and co-workers for efforts
to demonstrate binding of the peptide to DNA by NMR techniques. We
gratefully acknowledge technical assistance by H.
Büngen and U. Kaysser.
©1995 by The American Society for Biochemistry and Molecular Biology, Inc.
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  M. Reuter, J. Schneider-Mergener, D. Kupper, A. Meisel, P. Mackeldanz, D. H. Kruger, and C. Schroeder Regions of Endonuclease EcoRII Involved in DNA Target Recognition Identified by Membrane-bound Peptide Repertoires J. Biol. Chem., February 19, 1999; 274(8): 5213 - 5221. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| All ASBMB Journals | Molecular and Cellular Proteomics |
| Journal of Lipid Research | ASBMB Today |