ABSTRACT
To characterize the interaction between the homing endonuclease PI-
Sce
I and DNA, we prepared different DNA substrates containing the natural
recognition sequence or parts thereof. Depending on the nature of the
substrates, efficient cleavage is observed with a DNA containing
~
30 bp of the natural recognition sequence using supercoiled plasmids,
~
40-50 bp using linearized plasmids and
>
50 bp using synthetic double- stranded oligodeoxynucleotides. Cleavage of supercoiled plasmids occurs without accumulation of the nicked intermediate. In the
presence of Mn
2+
, DNA cleavage by PI-
Sce
I is more efficient than with Mg
2+
and already occurs with substrates containing a shorter part of the recognition
sequence. The requirements for strong binding are less stringent: a 35 bp
oligodeoxynucleotide which is not cleaved is bound as firmly as other longer
oligodeoxynucleotides. PI-
Sce
I binds with high affinity to one of its cleavage products, a finding which may
explain why PI-
Sce
I hardly shows enzymatic turnover
in vitro
. Upon binding, two complexes are formed, which differ in the degree of bending
(45
o
versus 75
o
). According to a phasing analysis bending is directed into the major groove.
Strong binding, not, however, cleavage is also observed with the genetically
engineered enzymatically inactive variant comprising amino acids 1-277. Models for binding and cleavage of DNA by PI-
Sce
I are discussed based on these results.
Homing endonucleases have been found in eubacteria, archea and eukaryotes. In
general they are coded for by open reading frames which are located in introns
or constitute inteins. Homing endonucleases, like restriction endonucleases,
are highly specific enzymes which cleave double-stranded DNA in the presence of Mg
2+
to produce 5'- or 3'-overhangs, with a 5'-phosphate and a 3'-OH. Unlike type II
restriction endonucleases they recognize extraordinarily long sequences, >15 bp in length, with no apparent symmetry. Recognition, however, is not as stringent as that of restriction
endonucleases, because in general, homing endonucleases are tolerant to the
substitution of individual base pairs. The cleavage of DNA by homing
endonucleases
in vivo
is likely to be followed by a double strand break repair accompanied by a
recombination event in which the coding sequence for the homing nuclease is
`inserted' into an allele that lacks it, a process which is called `homing' and
has given this class of enzymes their name (reviewed in
1
-
3
). As extremely rare cutting endonucleases they could become very useful for
engineering changes into genomes (
4
).
Homing endonucleases are a heterogeneous class of enzymes. Depending on
characteristic sequence motifs, different families are distinguished. The
largest family is characterized by the LAGLIDADG motif, which is present in the
first homing endonuclease that was identified, I-
Sce
I (
5
), but is also observed in other enzymes interacting with nucleic acids, e.g.
maturases (
6
). The LAGLIDADG motif is likely to be involved in catalysis (
7
,
8
), and in most homing endonucleases occurs in two copies. One of the best
characterized members of the LAGLIDADG family of homing endonucleases is PI-
Sce
I, formerly known as VDE (
PI-
Sce
I was purified to homogeneity from its natural host, as well as in recombinant
form from
Escherichia coli
(
11
). Both preparations were indistinguishable with respect to behaviour during
purification, apparent molecular weight (50 kDa), immunological reactivity and,
in particular, catalytic properties. Like all homing endonucleases, PI-
Sce
I requires Mg
2+
ions for DNA cleavage, which can be replaced by Mn
2+
, but not, however, Co
2+
, Ca
2+
or Zn
2+
(
11
). The minimal nucleotide sequence for cleavage was delimited by a primer
extension method using the natural
VMA1
[Delta]
vde
DNA as the template to a 30 bp sequence. In DNase I footprinting experiments a
35 bp region was found to be protected. A hydroxyl radical protection analysis
showed that within this region PI-
Sce
I afforded protection more on the upper than on the lower strand and more on the
right side of the cleavage site than on the left, suggesting that contacts
between the protein and the DNA are highly asymmetric. A 67 bp DNA comprising
the region protected against DNase I is bound by wild-type PI-
Sce
I in the absence of Mg
2+
with an apparent
K
Diss
of 3.7 nM, as determined by an electrophoretic mobility shift analysis. Similar
affinities were measured for several D -> A, N or E mutants at positions 218 and 326, i.e. within the first and
second LAGLIDADG motif. These mutants are completely (D -> A or N) or largely (D -> E) inactive, even in cleaving a substrate nicked specifically in one
or the other strand (
8
).
One of the most interesting questions regarding the mechanism of DNA recognition
and cleavage by PI-
Sce
I as well as by other homing nucleases of the LAGLIDADG family, concerns the question of how specific binding triggers phosphodiester bond cleavage in the two
strands of the double-stranded substrate. To address this problem we have begun to study the
kinetics of the reaction and to resolve individual steps using different
substrates and enzyme variants. In pursuit of this approach we noticed that PI-
Sce
I has a more extended recognition site than suggested on the basis of the primer
extension experiments mentioned above. Here, we present results which
demonstrate that PI-
Sce
I depending on the substrate and the divalent metal ion cofactor requires >50 bp
of its natural substrate for cleavage to occur, that the DNA is bent by 75o into the major groove and that one part of the DNA cleaved remains firmly
associated with the enzyme and, thereby, prevents efficient turnover. It is
shown here, furthermore, that supercoiled DNA is cleaved without accumulation
of a nicked intermediate, which could be interpreted to mean either that two
catalytic centres are involved in double strand cleavage in a concerted
reaction, or only one catalytic centre, with the second cleavage occuring
faster than the first one and within one binding event.
DNA from
Saccharomyces cerevisiae
(commercial baker's yeast from Deutsche Hefewerke) was prepared as described (
12
). Using this DNA a 1391 bp DNA fragment containing the PI-
Sce
I gene was obtained by PCR with the following protocol: 100 ng template, 0.4 [mu]M of each primer (5'-GCGGATCCGCATGCTTTGCCAAGGGTACCAATG-3' and 5'-CCGGCGTCGACGTCAGCAATTATGGACGACAACCTGG-3'), 0.2 mM dNTPs
(Pharmacia), 10 mM Tris-HCl, pH 8.0, 50 mM KCl, 1.5 mM MgCl
2
, 0.1 mg/ml gelatine, 2 U
Taq
polymerase (Amersham) in a reaction volume of 50 [mu]l (cycle 1, 390 s at 95oC; cycle 2, 90 s at 95oC, 90 s at 55oC and 90 s at 72oC; cycle 2 was repeated 30 times; cycle 32, 360 s at 72oC and 30 s at 0oC). After digestion with
Bam
HI (USB) and
Sal
I (AGS, Heidelberg, Germany) spin column purified PCR products were ligated into
pHisRV (
13
), cleaved with the same enzymes, to give pHisPI-
Sce
I. This plasmid codes for PI-
Sce
I with an additional N-terminal affinity-tag, Met(His)
6
GlySerAla. By DNA sequencing we detected several modifications in the nucleotide
sequence compared to the sequence given by Hirata
et al.
(
9
). Based on the DNA sequence the protein sequence of the PI-
Sce
I obtained differs in three positions (R44S, V67M, I132V). This difference seems
to be a characteristic of the particular yeast strain used, as it was found in
several independent PCR cloning experiments.
The plasmid pHisPI-
Sce
IN coding for a truncated form of the PI-
Sce
I consisting of amino acids 1-277 was prepared by PCR. The plasmid pHisPI-
Sce
I was used as template in a PCR reaction with
Pfu
-DNA-polymerase (Stratagene) and the primers 5'-GCGGATCCGCATGCTTTGCCAAGGGTACCAATG-3' and 5'-CCGGCGTCGACGTCAACGTATACCATTACCTCTGAC-3', carried out
essentially as described for the gene of the full length PI-
Sce
I. The purified PCR product was cleaved with
Bam
HI and
Sal
I and inserted into the plasmid pHisPI-
Sce
I which was digested before with the same restriction enzymes, to give pHisPI-
Sce
IN. The sequence of the gene coding for the N-terminus was confirmed by sequencing.
For overexpression of His
6
-tagged PI-
Sce
I,
Escherichia coli
strain LK111([lambda]) was used. LK111([lambda]) cells harbouring the plasmid pHisPI-
Sce
I in which the PI-
Sce
I gene is under the control of the
tac
promoter were grown at 25oC in 500 ml LB-medium containing 75 [mu]g/ml ampicillin in a 1l shaking flask to an absorbance of 1
A
600
(1 cm). After induction with 1 mM IPTG, the culture flask was incubated
overnight at 25oC. Cells were harvested by centrifugation and washed once with STE buffer
[10 mM Tris-HCl, pH 8.0, 0.1 mM EDTA and 0.1 M NaCl]. The cell pellet was resuspended
in 40 ml lysis buffer [10 mM Tris-HCl, pH 8.0, 200 mM KCl, 15 mM imidazole, 0.1 mM DTE, 5% (v/v) glycerol].
After sonication (5 * 1 min at 4oC), the cell debris was removed by centrifugation. The following
steps were performed at 4oC. The supernatant was applied to 2.5 ml (bed volume) of Ni-NTA-resin (QIAGEN) equilibrated with lysis buffer. The column was washed
with 50 ml of the same buffer and PI-
Sce
I was eluted with 7 * 1ml of elution buffer [10 mM Tris-HCl, pH 8.0, 200 mM KCl, 200 mM imidazole, 0.1 mM DTE, 5% (v/v)
glycerol]. Individual fractions were analyzed by SDS-PAGE on a 10% (w/v)
polyacrylamide gel. PI-
Sce
I containing fractions were pooled, diluted 5-fold with PDL buffer [30 mM KPi, pH 7.2, 0.1 mM DTE, 0.01% (w/v) Lubrol
PX] supplemented with 0.1 M NaCl and loaded onto 2.5 ml (bed volume) of
phosphocellulose P11 (Whatman) equilibrated with PDL buffer, 0.1 M NaCl. The
column was washed with 25 ml of PDL buffer, 0.1 M NaCl and the protein was
eluted with 15 * 1 ml of PDL buffer, 2 M NaCl. PI-
Sce
I containing fractions were pooled and dialyzed overnight against storage buffer
[10 mM KPi, pH 7.4, 200 mM KCl, 0.1 mM EDTA, 1 mM DTE, 50% (v/v) glycerol]. The
PI-
Sce
I concentration was determined using the extinction coefficient of 4.68 * 10
4
/M/cm (
8
). The yield of >= 95% pure PI-
Sce
I usually was 2-2.5 mg/l culture. This preparation had the same specific activity as
authentic PI-
Sce
I isolated from yeast (
11
).
For the preparation of the N-terminal PI-
Sce
I fragment
E.coli
cells harbouring pHisPI-
SceI
N were fermented as described above. Purification of the protein was performed
under denaturing conditions, because the protein forms insoluble aggregates in
the
E.coli
cells. After disintegration of
E.coli
cells the N-terminal PI-
Sce
I fragment was found in the pellet together with the cell debris. This pellet
was resuspended in 20 ml of buffer A [lysis buffer, 6 M urea] by shaking
overnight. Insoluble debris was removed by centrifugation and the supernatant
was applied to a Ni-NTA resin column equilibrated with buffer A. After washing
with 50 ml of buffer A the protein was eluted with 7 * 1 ml of buffer B [elution buffer, 6 M urea], fractions containing the N-terminal PI-
Sce
I fragment were combined and dialyzed thoroughly against storage buffer. Any
precipitate formed was removed by centrifugation. The protein concentration was
determined from the absorbance at 280 nm. The extinction coefficient of the N-terminal fragment was calculated on the basis of its content of aromatic
amino acids (
14
) to be [epsilon]
280 nm
= 3.1 * 10
4
/M/cm. The yield of soluble truncated protein was 100-150 [mu]g/l culture.
Oligodeoxynucleotides were synthesized on solid support with a Milligen cyclone
plus DNA synthesizer using [beta]-cyanoethylphosphoramidites obtained from PerSeptive Biosystems.
Individual oligodeoxynucleotides were purified by gel electrophoresis. Eight
different double-stranded oligodeoxynucleotides were used as substrates in binding (B-G) and six oligodeoxynucleotides in cleavage reactions with PI-
Sce
I (B-F and G; see Fig.
1
). If necessary, annealed oligodeoxynucleotides were labelled using [[alpha]-
32
P]dATP (Amersham) and Klenow enzyme (AGS). Oligodeoxynucleotide F
II
was labelled using [[gamma]-
32
P]ATP (Amersham) and T4 polynucleotide kinase (MBI Fermentas).
A 311 bp DNA fragment carrying the PI-
Sce
I cleavage site in the centre was generated by PCR using
Xmn
I (AGS) linearized pBSVDEX plasmid (New England Biolabs) as template. Template
DNA (50 ng) was used in a 100 [mu]l volume containing 0.4 [mu]M of each primer (5'-GCGCGGATCCAGGTCAAAGAGTTTTGG-3' and 5'-GCGTCGGATCCAAGCTTCTCTGGCTGC-3') and 0.2 mM dNTPs
and 10 [mu]Ci of [[alpha]-
32
P]dATP (Amersham). The reaction was carried out with 5 U
Taq
polymerase (Promega) in
Taq
1* reaction buffer (cycle 1, 300 s at 95oC; cycle 2, 60 s at 54oC, 60 s at 72oC and 30 s at 95oC; cycle 2 was repeated 30 times; cycle 32, 60 s at
54oC and 300 s at 72oC). The PCR product was purified by QIAquick PCR purification kit
(QIAGEN).
Oligodeoxynucleotides A-F and G (Fig.
1
) were ligated into the vector pAT153 which had been cleaved with
Bam
HI to produce the plasmids pAT-A to pAT-G. Plasmid pBSVDEX (
11
) was obtained from New England Biolabs. After transformation into
E.coli
DH5[alpha] and fermentation,
plasmid DNA was prepared using preparation kits (QIAGEN) according to the
instructions of the supplier.
For electrophoretic mobility shift assays, PI-
Sce
I was mixed with binding buffer [10 mM Tris-HCl, pH 7.5, 1 mM EDTA, 50 mM NaCl, 0.05% (w/v) nonfat dry milk, 5% (v/v)
glycerol, 10 mM DTT] in a total volume of 10 [mu]l containing in addition 0.1 [mu]g of the nonspecific DNA carrier poly(dI-dC) (Pharmacia), and 10 000 c.p.m. of
32
P-labelled DNA probe. The mixture was incubated for 60 min at room
temperature. To stop the reaction 3 [mu]l of loading buffer [10% (w/v) Ficoll, 15% (v/v) glycerol, 50% binding
buffer, 40 mM EDTA, 0.2% (w/v) bromophenol blue, 0.2% (w/v) xylene cyanol] were
added. Samples were loaded onto a native 7% (w/v) polyacrylamide
(acrylamide:bisacrylamide, 19:1) minigel containing 0.5* TBE [44.5 mM Tris-borate, pH 8.4, 1 mM EDTA]. Gels were prerun at 10 V/cm for 30 min
and after loading run for an additional 2-4 h at room temperature. Following electrophoresis radioactive bands were
detected and quantified using an Instant Imagertm system (Canberra Packard). For the determination of binding constants, PI-
Sce
I concentrations were varied from 1 to 250 nM, at a constant concentration of 1 nM (oligodeoxynucleotide F and F
II
) or 4 nM (311 bp shift DNA) of
32
P-labelled DNA probe.
Cleavage reactions were performed with oligodeoxynucleotides (1 [mu]M) varying in length between 35 and 61 specific bp (oligodeoxynucleotides B-F and G, Fig.
1
), a
32
P-labelled 311mer PCR-generated fragment with a central PI-
Sce
I cleavage site (4 nM) and with supercoiled or
Xmn
I-linearized plasmids pAT-A to pAT-G, and pBSVDEX (8 nM) as substrates in 10 [mu]l cleavage buffer [10 mM Tris-HCl, pH 8.5, 100 mM KCl, 1 mM DTT, 100 [mu]g/ml BSA] containing 2.5 mM EDTA or 2.5 mM MgCl
2
or 2.5 mM MnCl
2
, respectively, for 1.5 h at 37oC. Concentrations of PI-
Sce
I were varied as indicated. The reactions were terminated by the addition of 3 [mu]l of stop buffer [100 mM EDTA, pH 8.0, 25% Ficoll, 0.1% (w/v)
bromphenolblue, 0.1% xylene cyanol]. The progress of the cleavage reaction was
analyzed by electrophoresis in 1* TPE buffer [80 mM Tris-phosphate, pH 8.2, 2 mM EDTA] either on a 1% (w/v) agarose gel or
on a 7% (w/v) or 12% (w/v) polyacrylamide gel, which were analyzed using an
Instant Imagertm system (Canberra Packard) or after ethidium bromide staining using the
Intas gel documentation system (Intas, Göttingen, Germany).
The substrates for the circular permutation assay were prepared using pBend2VMA1[Delta]vde. This plasmid was prepared by insertion of the 76 bp
Eco
RI-
Sca
I fragment derived from pBSVDEX containing the PI-
Sce
I homing site into the
Sal
I site of the plasmid pBend2 (
15
) after the resulting termini had been converted to blunt ends by incubation
with Klenow fragment. The plasmid pBend2VMA1[Delta]vde was used as a template in a PCR, essentially as described before,
with the primers 5'-GAGGCCCTTTCGTCTTCAAGAATTC-3' and 5'-GTGATAAACTACCGCATTAAAGCTT-3' which are complementary to
the sequences flanking the pBend2VMA1[Delta]vde multiple cloning site. The PCR was carried out in the presence of [[alpha]-
32
P]dATP to label the PCR products. The 201 bp bending probes were generated by
restriction enzyme digestion of the PCR product with the enzymes indicated in
Figure
6
and purified using 6% (w/v) polyacrylamide gels. Electrophoretic mobility shift
assays were performed on 6% (w/v) polyacryamide gels (acrylamide:bisacrylamide, 29:1) under conditions as described above. To calculate the centre of the bending, the relative
mobility of each protein-DNA complex was plotted as a function of the distance of the middle of
the cleavage site to the ends of the bending probe. A second order polynomial
curve was fitted to the data. The bend centre was estimated from the position
at which this curve reached a minimum. The DNA flexure angle was calculated
from the ratio of the mobilities of the slowest and the fastest migrating
complex ([mu]M/[mu]E) which was fitted to the equation [mu]M/[mu]E = cos([alpha]/2), where [alpha] is the bend angle (
15
). The results were standardized by a circular permutation analysis using pBend2
fragments complexed with
Eco
RV (
16
).
The plasmid construct for the phasing analysis was kindly provided by Dr R.
Niedenthal (
17
) containing a sequence with three phased A-tracts with an intrinsic bend of 54o (
18
) which are separated by a linker of variable length (0, 2, 4, 6, 8, 10 bp) from
the
Eco
RV restriction site in the multiple cloning site of pBluescript II SK. A
Hin
dIII-
Sca
I fragment derived from pBSVDEX (
11
) containing the PI-
Sce
I homing site was inserted into the
Hin
dIII-
Eco
RV site of the plasmids. These constructs were used as templates in a PCR with
the primers T3 and T7 (Stratagene). The PCR was carried out in the presence of
[[alpha]-
32
P]dATP. The resulting phasing probes contain the PI-
Sce
I recognition site and, separated by spacers of different length, which varied
over one turn of the DNA helix, form an intrinsic DNA bend induced by the A-tracts.
Based on a primer extension analysis it was reported that PI-
Sce
I recognizes a 30 bp sequence (
11
) in the context of the
VMA1
[Delta]
vde
sequence which is the target into which the coding sequence of the PI-
Sce
I
in vivo
is inserted after cleavage by PI-
Sce
I (
19
). Parts of the
VMA1
[Delta]
vde
sequence are shown in Figure
1
. The cleavage and the insertion sites are indicated, as well as the minimal
recognition site and the boundaries defined by a DNase I footprint analysis (
8
). As we had observed that PI-
Sce
I does not cleave oligodeoxynucleotide A (Fig.
1
) which corresponds to the presumptive minimal recognition site as defined by
Gimble and Thorner (
11
), we have compared the ability of PI-
Sce
I to cut DNA substrates containing a more or less extended part of the natural
VMA1
[Delta]
vde
sequence.
For this purpose, double-stranded oligodeoxynucleotides varying between 35 and 61 specific bp in
length and having 5' overhangs of 4 nt (Fig.
1
) were synthesized. As these oligodeoxynucleotides were not (A-E) or only slowly cleaved (F, G) by PI-
Sce
I (data not shown), these oligodeoxynucleotides were cloned into pAT153, to give
pAT-A to pAT-G. These plasmids together with a reference plasmid, pBSVDEX (
11
), were used to try to define the recognition site of PI-
Sce
I on polynucleotide substrates. Cleavage experiments were carried out with
Xmn
I linearized plasmids as well as with supercoiled plasmids (not, however, with
relaxed circular DNA), both in the presence of Mg
2+
or Mn
2+
. The results are given in Table
1
and can be summarized as follows. All plasmid substrates are cleaved in the
presence of Mn
2+
with similar rates, the supercoiled form being cleaved by a factor of 1.4-3.7 more rapidly than the linear form. In the presence of Mg
2+
which can be considered the natural divalent metal ion cofactor, pAT-A to pAT-B are only cleaved in the supercoiled form, not, however, to a
detectable extent in the linear form. All other plasmid DNA substrates are
cleaved also in the linear form, in general by a factor of 0.7-2.8 more slowly than in the supercoiled form. From these data it can be
concluded that the definition of the recognition site for PI-
Sce
I depends on the kind of substrate used (oligodeoxynucleotide versus
polynucleotide, linear versus supercoiled form), as well as on the divalent
metal cofactor (Mg
2+
versus Mn
2+
) present in the assay mixture.
It is noteworthy that cleavage of supercoiled DNA substrates proceeds without accumulation of
the open circular intermediate, neither in the presence of Mg
2+
nor in the presence of Mn
2+
(Fig.
2
). This finding demonstrates that the two phosphodiester bond cleavage events
occur in a concerted reaction.
As can be deduced from published specific activity data PI-
Sce
I is a very slow enzyme: 0.1 moles of sites are cleaved per mole PI-
Sce
I per hour (
11
). Figure
3
shows that pBSVDEX which contains a 781 bp segment of the
VMA1
[Delta]
vde
sequence, as well as a 311 bp PCR fragment derived from pBSVDEX, are cleaved
only with at least stoichiometric amounts of enzyme and require incubation
times of >1 h. A similar result was obtained with other preparations as well as
with a commercial preparation of PI-
Sce
I (New England Biolabs). These data suggest that the enzyme turns over only very
slowly. As will be shown below, this is probably due to the fact that PI-
Sce
I remains firmly bound to one of its cleavage products.
To find out why PI-
Sce
I does not cleave short oligodeoxynucleotide substrates and does not turn over
efficiently, the binding of PI-
Sce
I to oligodeoxynucleotides B-F and G was analyzed and compared to the binding of
PI-
Sce
I to the 311 bp PCR fragment. These binding experiments were carried out in the
absence of Mg
2+
to prevent cleavage. In Figure
4
an electrophoretic mobility shift experiment with 4 nM of the 311 bp PCR
fragment in the presence of increasing amounts of PI-
Sce
I is shown. The complex observed with enzyme concentrations up to 100 nM is
specific, as it cannot be competed by 20 [mu]g/ml poly(dI-dC), but effectively by the unlabelled 311 bp PCR fragment
(data not shown). These results show that PI-
Sce
I binds specifically to DNA under these conditions. At concentrations of PI-
Sce
I of >100 nM an additional complex with slow mobility is observed which is non-specific, as its formation can be suppressed by 20 [mu]g/ml poly(dI-dC). According to the evaluation of several
independent titrations the apparent equilibrium constant for the formation of
the specific complex is 7.5 +- 2.5 nM. Very similar affinities were determined for the
oligodeoxynucleotides B-F and G. This finding suggests that the inability of
oligodeoxynucleotides to function as good substrates for PI-
Sce
I is not due to a low affinity.
Table 1
The DNaseI footprint of PI-
Sce
I on its DNA substrate extends over 35 bp (
8
) and as shown above, under certain conditions an even longer sequence is needed
for cleavage to occur. The recognition of such a long stretch of DNA could be
achieved by a distortion of the DNA to enlarge the contact area between
substrate and enzyme, similarly as shown recently for I-
Tev
I, I-
Tev
II and I-
Ppo
I (
21
-
23
). For the conformational analysis of the PI-
Sce
I-substrate complex an electrophoretic mobility shift assay was performed using
DNA fragments which contain a circular permuted recognition site. A bending of
the DNA in the protein-DNA complex would lead to a substantial decrease in the electrophoretic
mobility of the protein-DNA complex. This effect depends on the degree of the bend and the
position of the distortion in the fragment (
24
). It is more pronounced, if the position of the bend is at the centre of the
DNA fragment. To produce the bending vectors, a 76 bp fragment containing the
PI-
Sce
I recognition site was inserted into the plasmid pBend2 (
15
). This plasmid allows one to generate bending probes by restriction enzyme
digestion that are identical in size but differ in the position of the PI-
Sce
I binding site within the probe (Fig.
6
A). The electrophoretic mobility shifts obtained with PI-
Sce
I and these probes are shown in Figure
6
B. The unbound DNA fragments display a nearly uniform electrophoretic mobility
regardless of the position of the PI-
Sce
I binding site within the probe. Bending probes which were incubated with PI-
Sce
I show two additional bands. The upper protein-DNA complexes (UC) display a pronounced position-dependent variation in electrophoretic mobility with the slowest
migration when the PI-
Sce
I recognition site is at the centre of the probe, whereas the lower complexes
(LC) display only a slight variation in electrophoretic mobility. The relative
mobility of each protein-DNA complex was plotted as a function of the distance of the cleavage
site to the ends of the bending probe (Fig.
6
C). The centre of the bend was estimated from the minimum of the curve and
mapped to position 7 +- 1 bp to the right of the middle of the PI-
Sce
I cleavage site. The flexure angles of the protein-DNA complexes were calculated to be 75o (+-15o) in the upper complex and 45o (+-15o) in the lower complex (based on eight independent
experiments). It is important to note, that according to the electrophoretic analysis under denaturing
conditions, the DNA in both the upper and lower complex is not nicked (data not
shown).
We have shown here that the recognition site for the homing endonuclease PI-
Sce
I is not as clearly defined as for example for type II restriction endonucleases
(
29
). Depending on the kind of substrate, viz. oligodeoxynucleotide or plasmid DNA,
in a relaxed or supercoiled form, and depending on the divalent metal ion
cofactor used, viz. Mg
2+
or Mn
2+
, a more or less extended sequence is required for efficient cleavage. In the
presence of Mg
2+
and with a macromolecular DNA which is not under torsional stress, a sequence
of ~40 bp is needed for efficient cleavage.
The requirements for strong binding are less stringent than for efficient
cleavage. Oligodeoxynucleotides comprising ~15 bp to the left and right of the homing site are not cleaved by PI-
Sce
I but are bound as well as oligodeoxynucleotides which are longer by 20 bp and
are cleaved by PI-
Sce
I. In this context it should be emphasized that strong and specific binding is
not dependent on divalent metal ions, while cleavage is (
8
,
11
). The fact that the structural requirements for specific cleavage are more
stringent than for strong binding was not unexpected, as recognition is a
dynamic process which also includes interactions formed in the transition state
of the enzymatic reaction. Such interactions can be decisive, in particular
when major conformational rearrangements in the protein and in the DNA take
place after initial complex formation. A very extreme example is provided by
the restriction endonuclease
Eco
RV which in the absence of Mg
2+
binds only non-specifically to DNA (
31
), but in the presence of Mg
2+
, however, discriminates very effectively between specific and non-specific sites (
32
,
33
). While the DNA is not bent in the nonspecific complex (
34
), it is bent in the specific complex in the presence of divalent cations (
16
,
35
).
A substantial part of the affinity of PI-
Sce
I to its substrate must be provided by interactions between the 3'-half of the recognition site and the N-terminal half of PI-
Sce
I, as demonstrated by the fact that one of the products of the reaction is bound
even more firmly to the enzyme than the corresponding substrate and that a
recombinant PI-
Sce
I fragment comprising amino acids 1-277 including one LAGLIDADG motif (amino acids 210-219) binds specifically to but does not cleave DNA containing the recognition sequence. It remains to be seen whether the C-terminal half of PI-
Sce
I also interacts with DNA; so far we have not been able to produce in soluble
form C-terminal fragments of PI-
Sce
I containing the other LAGLIDADG motif (amino acids 318-326).
The firm binding of one of the products of the PI-
Sce
I catalyzed cleavage of DNA to the enzyme precludes efficient turnover. In this
respect, PI-
Sce
I resembles I-
Sce
I which also belongs to the LAGLIDADG family of homing endonucleases. I-
Sce
I binds strongly to the 3'-half of its recognition site and in steady-state cleavage experiments exhibits two step kinetics, with a
burst whose magnitude depends on the enzyme concentration and a very slow
turnover limited by product release (
20
). Such a behaviour, however, is not the rule for the LAGLIDADG type homing
endonucleases, as it was shown that I-
Por
I readily turns over and does not bind to either of its products (
36
). Strong binding to one of the products is also seen with homing endonucleases
belonging to other families, e.g. I-
Tev
II (
22
). On the other hand, I-
Ppo
I, also not a member of the LAGLIDADG family, shows no affinity towards its
products (
23
). Thus, there is no correlation between family membership on one side and
effective turnover on the other side.
We thank Dr A. Jeltsch for critical reading of the manuscript and Sonja Franke
for technical assistance. This work has been supported by grants from the Fonds
der Chemischen Industrie and the Bundesministerium für Bildung, Wissenschaft, Forschung und Technologie.
Plasmid
Length of original
Form
Activity * 10
-3
sequence (bp)
[M DNA cleaved/min * M enzyme]
a
Mg
2+
Mn
2+
pAT-A
31
linear
n.d.c.
b
5
supercoiled
3.0 (+- 0.1)
18.7
pAT-B
35
linear
n.d.c.
b
6.7
supercoiled
3.2
23
pAT-C
40
linear
1.5
n.d.
supercoiled
1.0 (+- 0.1)
n.d.
pAT-D
40
linear
2.3
n.d.
supercoiled
6.0
n.d.
pAT-E
51
linear
3.0 (+- 0.1)
13.7
supercoiled
8.4
30.9
pAT-F
56
linear
3.0 (+- 0.1)
9.8 (+- 0.1)
supercoiled
5
13.4
pAT-G
61
linear
3.4 (+- 0.3)
13.0 (+- 0.3)
supercoiled
6.2
20.2
pBSVDEX
758
linear
3.2 (+- 0.2)
12.3 (+- 0.1)
supercoiled
5.6
20.3
REFERENCES
Return