Type I enzymes are complex multifunctional proteins which cleave unmodified DNA in the presence of SAM, ATP and Mg2+. If hemimethylated DNA is the substrate then the enzyme is converted into its modifying form and methylates the second strand of the DNA. Fully modified DNA is resistant to DNA cleavage and the enzyme dissociates from it. SAM is thought to act as an allosteric effector, at least for R.EcoKI, which converts the enzyme into an active form which is capable of specific DNA binding. ATP is known to be involved in the determination of the methylation status of the DNA. The enzyme reaction mechanism is complicated and results in DNA cleavage at random sites distant to the recognition site. One model for DNA cleavage has shown how this can result from a cooperative reaction in which two endonuclease molecules, bound to neighbouring recognition sites, meet following DNA translocation and then cleave the DNA. Translocation of the DNA requires ATP and the endonuclease is a powerful ATPase. Interestingly, the endonuclease does not turn over following DNA cleavage and appears to remain bound to the reaction product. This DNA-protein complex appears to be responsible for the massive ATP-hydrolysis continuing for a long period after DNA cleavage.
Electron Micrograph of the EcoR124I
endonuclease
bound to a two-site plasmid DNA molecule
We have described a two plasmid system which allows
overproduction of the R.EcoR124I restriction endonuclease. The
endonuclease has been purified to homogeneity in milligram amounts and has been
shown to be fully active for both restriction and modification. Unexpectedly,
the enzyme was found to require only ATP and Mg2+ for ATPase activity
and DNA cleavage; S-adenosyl methionine, (SAM) which has previously been
described as a cofactor of Type I restriction enzymes, is NOT required by R.EcoR124I.
However, SAM was found to stimulate the rate of ATPase
activity and DNA cleavage Janscak
(1996). This may occur through an increase in specific DNA binding by R.EcoR124I
in the presence of SAM as indicated by our surface plasmon
resonance experiments. These functional differences from the well described
R.EcoKI restriction endonuclease are reflected in a possible structural
difference between the two enzymes; namely that the
stoichiometry of R.EcoR124I
was first reported to be R1M2S1 (Janscak
et al 1996), while that of R.EcoKI
is R2M2S1. However,
this initial description of the unusual stoichiometry for EcoR124I was later
shown to actually reflect an unusual situation where dissociation of the enzyme
occurs resulting in a mixture of the R2-complex and the R1-complex was present
in the cell (Janscak et al 1998).
The R1-complex was found to be restriction-deficient, but capable of
unidirectional translocation (Janscak
et al 1998; Seidel et al 2004;
2005).
The dissociation of The EcoR124I endonuclease has introduced the concept that there may be fundamental differences between the native holoenzyme isolated from E. coli [R124] and the in vitro assembled enzyme, produced by mixing purified MTase and HsdR (Bianco & Hurley, 2005). However, all of our evidence suggests that both enzymes behave the same way, but that during any functional analysis of the holoenzyme the concentration of the enzyme used is very important as dissociation will occur (Figure to right).
The cleavage of linear DNA substrates is a random process in which the
specifically bound endonuclease translocates DNA until it meets another enzyme.
This collision then promotes cleavage of the substrate
(Studier & Bandyopadhyay, 1988).
This model was suggested following studies of cleavage of T7 DNA by EcoKI,
but does not really explain what happens when only one recognition site is
present on the DNA, it also does not really address the situation with
supercoiled substrates. In the case of a supercoiled substrate
increased degree of supercoiling has been
shown to inhibit restriction (Janscak
1996)which suggests that the process of translocation may be inhibited by
supercoiling. In collaboration with
Prof. Halford's group at
the University of Bristol, we have studied the cleavage of supercoiled DNA with
plasmids carrying one or two sR124I recognition sites. Both
substrates are cleaved by the same mechanism inferring
that the model applied to linear DNA is incorrect for circular DNA
Szczelkun (1996).
Supercoiled DNA, when translocated through the specifically bound enzyme, will
be over-wound on one side of the complex and under-wound on the other side. This
means that the two bound enzymes, on a two-site plasmid, cannot meet and that
some other mechanism must ensure cleavage occurs. We suggest that when all of
the available DNA has been translocated, which may mean when the DNA topology
prevents further translocation, then the translocation process will stop and DNA
cleavage will occur. Therefore, any process which stalls the DNA translocation
will lead to cleavage - on linear DNA this will occur when two enzymes collide -
on circular DNA this will occur when translocation is sufficiently hindered by
the DNA supercoiling that the process stalls.
We have also investigated the nature of the attachment of the enzyme to DNA prior to the translocation. In order to determine that the DNA can be bound across a distance (in trans), or whether the adjacent DNA is bound, we have used a catanene carrying a single sR124I site. If the DNA is bound at a second site in trans then both catanene molecules should be cut. However, this was not the case indicating that, following specific binding at sR124I the second binding is non-specific and in cis. Nicked-circle DNA is an intermediate of cleavage reaction. This may be required to allow the initial stage of translocation, a loop will have to be produced following the non-specific binding, and for translocation of this loop to occur the DNA would have to be relaxed.
We have cloned the hsdR gene of the EcoR124I
restriction endonuclease, confirming the presence of an independent promoter for
this gene by means of a complementation assay with MTase
(Zinkevich 1997). In
addition, we have also cloned the hsdR gene into the expression vector
pTrc99A. This has allowed us to purify the subunit in milligram quantities and
to study the enzymatic properties of the individual subunit. The presence of a
Walker Type I ATP-binding site within the HsdR subunit suggested that the
subunit might be capable of independent enzymatic activity. The purified HsdR
subunit was found to be a soluble monomeric protein capable of a DNA and Mg2+-dependent
ATP hydrolysis. The subunit was found to have a weak nuclease activity both
in vivo and in vitro, and to bind plasmid DNA; although was not
capable of binding a DNA oligoduplex. We were also able to reconstitute the
fully active endonuclease from purified M.EcoR124I and HsdR. This is the
first clear demonstration that the HsdR subunit of a Type I restriction
endonuclease is capable of independent enzyme activity, and suggests a mechanism
for the evolution of the endonuclease from the independent methylase by
acquisition of an additional subunit capable of nuclease activity.
The DNA binding was found to be unusual in that while it was weak (Kd = 0.2uM)
the binding was not non-specific. The subunit cannot bind an oligonucleotide
(30-mer) but was able to bind two HinfI fragments of the plasmid pDRM-1R
(a derivative of pTZ19R carrying a sR124 recognition site). Analysis
of these two DNA fragments failed to reveal a common sequence greater than 4-bp
in length suggesting the specificity of the binding may involve DNA curvature
rather than DNA sequence. The length of the fragment is not the sole determinant
for DNA binding since the digest used in the study had fragments present of
intermediate size to those retarded in the gel. The level of ATPase activity is
similar to that observed with Type III restriction endonucleases. This suggests
a "looping" model may be involved in the DNA translocation carried out by both
enzymes, perhaps with the DNA being wrapped around the HsdR subunit of the Type
I restriction endonuclease prior to the main translocation mechanism.
The hsdM gene was separately cloned into the vector pUC119 as part of the original isolation of the MTase (Patel 1992). The subunit also over-produced from the lac promoter of this plasmid allowing purification of the subunit. The protein was found to be a monomer of 58kDa and was found in the soluble fraction. While no individual enzymatic activity has been shown for HsdM, the subunit is responsible for binding the methyl donor SAM. Interestingly the subunit was found to be have very much as a globular protein, being quite resistant to proteolysis; although, a small tail of the protein was removed by trypsin.
The hsdS gene was cloned into the plasmid pTZ19R, a phagemid that allowed us to use site-directed mutagenesis to introduce a NdeI site at the start codon of the gene and to remove an internal NdeI site (Patel 1992). This allowed us to re-clone the gene into the powerful expression vector pET3a and thus overproduce the protein. The protein (46kDa) was found in the insoluble fraction, and despite all efforts to solubilise the protein we were unsuccessful. This insolubility of HsdS has more recently been used to identify regions of the protein involved in protein-protein interactions with HsdM (Abadjieva 1994). In the presence of HsdM a soluble methylase is produced.
That HsdS is responsible for DNA recognition has been clearly demonstrated by mutagenesis, including production of a deletion mutant with a novel DNA specificity based on dimerisation of individual "half" HsdS subunits (Abadjieva 1993). In order to be able to delineate more precisely the regions within the HsdS subunit that are responsible for protein-protein interactions, we have constructed a series of deletion mutants of the C-terminal region of the protein. We have previously shown that an in vivo complementation assay allows us to discriminate between two classes of mutation in the hsdS gene DNA binding mutants and those affected in protein-protein interactions. One deletion of the C-terminal one-third of the HsdS subunit is still able to produce an active methylase when co-produced with HsdM and the resulting methylase methylates at a novel DNA specificity (GAAn7TTC). The arrangement of this recognition sequence suggests that two deleted HsdS subunits are arranged in an inverted orientation each subunit recognizing one half of the wild-Type recognition sequence. Analysis of three deletion mutants, combined with information from a computer-aided analysis of the charge on the amino acids of HsdS and data from a series of limited proteolysis experiments with the methylase, indicate that the protein can be divided into two parts. The N-terminal region of HsdS can interact with HsdM to produce a novel methylase, and a 40 amino acid region at the C-terminal end of the central conserved region of HsdS is important for protein-protein interactions with HsdM which may help to stabilize this novel methylase.
The LEFT graph shows the stimulation by SAM (solid circles) of ATP hydrolysis. While the RIGHT graph shows the stimulation of DNA cleavage by SAM (solid circles).
The LEFT graph shows cleavage of a plasmid with two recognition sites. The RIGHT graph shows cleavage of a plasmid with a single recognition site.
The rising line represents DNA-binding, while the falling line represents release from the DNA following dilution. The release in the absence of SAM is bimodal indicating a mixture of specific and non-specific DNA binding.
With increased degree of negative supercoiling the concentration of linear DNA produced by cleavage with EcoR124I is reduced.
Last modified on
21 September 2011
© Dr Keith Firman
Author Dr Keith Firman.