©
1996 Oxford University Press
3590-3592 Footnote
Bae
I, another unusual
Bcg
Bae I, another unusual Bcg I-like restriction endonuclease
Lauren E.
Sears
,
Bing
Zhou
,
Jason M.
Aliotta
,
Richard D.
Morgan
and
Huimin
Kong*
New England Biolabs, Inc., 32 Tozer Road,
Beverly
, MA 01915,
USA
Received May 13, 1996;
Revised and Accepted August 5, 1996
ABSTRACT
Bcg
I and
Bcg
I-like restriction endonucleases have a very distinct characteristic which
causes them to differ from the other classified restriction enzymes; they all
cleave double-stranded DNA specifically on both sides of the recognition sequence to
excise a short DNA fragment including the recognition sites. Here we report a
new
Bcg
I-like restriction endonuclease,
Bae
I, isolated from
Bacillus sphaericus
. Like
Bcg
I,
Bae
I also cleaves double-stranded DNA on both strands upstream and downstream of its recognition
sequence (10/15)ACNNNNGTAYC(12/7). There are two dominant polypeptides in the
final preparation of
Bae
I with molecular masses of
~
80 and 55 kDa. Both are slightly larger than the two
Bcg
I subunits.
Bae
I requires both Mg
2+
and AdoMet to cleave DNA. Accompanying bilateral cleavage activity, the
heteromeric
Bae
I also has an
N
6-adenine methyltransferase activity which modifies the symmetrically
located adenines within its recognition sequence.
6
,
10
). Increasing numbers of restriction endonucleases that do not fit into the
conventional classification have been reported. Among these, one family of
restriction endonucleases is unique in that they all cleave double-stranded DNA on both sides of their recognition sequences to excise a
short fragment containing the recognition site. Typically they require Mg
2+
and
S
-adenosylmethionine for cleavage. Four examples of this new type of
restriction enzyme have been documented:
Bcg
I (
3
),
Bsp
24I (
1
),
Cje
I and
Cje
PI (
9
). The genes coding for
Bcg
I have been cloned and sequenced (
4
). The
Bcg
IA gene codes for a 71.6 kDa protein that resembles certain m
6
A-specific DNA-methyltransferases. The
Bcg
IB gene codes for a 39.2 kDa protein. Neither protein can cleave or modify DNA
by itself, but together they form a complex of composition of A
2
B that can do both (
4
).
Here we report a new
Bcg
I-like restriction endonuclease,
Bae
I, isolated from
Bacillus sphaericus
, NEB 659. The yield of this enzyme was at least 3200 U/g of cells in the crude
cell extract. The
Bae
I protein has been purified to near-homogeneity by phosphocellulose, heparin sepharose, Mono-Q (Pharmacia) and Heparin Tsk (Toso Haas) column chromatography.
There are two dominant polypeptides in the final preparation of
Bae
I with molecular masses of ~80 and 55 kDa (Fig.
1
A). Thus both the
Bae
I and
Bcg
I enzymes contain two protein subunits, however they differ in the sizes of
these subunits.
Figure 1
.
(
A
) Coomassie-stained 10-20% gradient SDS-PAGE of purified
Bae
I endonuclease (lane 1) with
Bcg
I as a control (lane 2). (
B
) 1% agarose gel showing the electrophoresis patterns of
Bae
I digested DNA. Lane 1, lambda DNA +
Hin
dIII, [Phi]174 +
Hae
III, size standard; lane 2, lambda DNA +
Bae
I; lane 3, adenovirus 2 DNA +
Bae
I. (
C
) Electrophoresis pattern of
Bae
I digested lambda DNA on a 3 % agarose gel (NuSieve, FMC). Lane 1, pBR322 +
Msp
I, size standard; lane 2, lambda DNA +
Bae
I. Arrow indicates the 33 bp fragment.
Like
Bcg
I,
Bae
I requires both Mg
2+
and
S
-adenosylmethionine for cleavage.
S
-adenosyl-l-homocysteine can also replace
S
-adenosylmethionine as a cleavage cofactor in
Bae
I digestion. In the case of
Bcg
I,
S
-adenosyl-l-homocysteine is unable to stimulate its cleavage activity (
3
). Complete digestion of lambda DNA and adenovirus 2 DNA was achieved by
Bae
I and resulted in 10 and five bands respectively (Fig.
1
B). Analysis of these DNA banding patterns resulting from
Bae
I-cleavage suggests that the enzyme recognizes the asymmetric non-contiguous sequence: 5'-ACNNNNGTAYC-3', or 5'-GRTACNNNNGT-3' on the
complementary strand (Y: pyrimidine, R: purine). The cleavage site of
Bae
I was determined by comparing dideoxy sequencing ladders with polymerized
extension products cleaved with
Bae
I as detailed previously (
2
,
3
,
5
). Two prominent DNA bands were observed following digestion with
Bae
I (Fig.
2
, lane -): one positioned 11 bases before the recognition sequence, and another
positioned 12 bases following it. The lane marked + in Figure
2
shows the result of treating the - lane products with the Klenow fragment of
Escherichia coli
DNA polymerase I. Both fragments were reduced by 5 nt, suggesting that
Bae
I makes a double-strand cleavage on both sides of its recognition sequence to produce two
five-base 3' extensions. To verify that cleavage occurs on both sides of the
recognition sequence, lambda DNA was digested with
Bae
I and then loaded to a 3% agarose gel. The expected 33 base pair (bp) fragments
were observed (Fig.
1
C). We conclude that
Bae
I cleaves double-stranded DNA on both strands upstream and downstream of its recognition
sequence (10/15)ACNNNNGTAYC(12/7).
Figure 2
.
The cleavage site of
Bae
I. Plasmid DNA that contains a
Bae
I recognition sequence and a synthetic primer were used in a sequencing reaction
based upon the dideoxynucleotide chain termination method. An additional
extension reaction was carried out in the presence of four deoxynucleotides and
[
33
P]dATP containing the same plasmid and primer. The labeled substrate was then
digested with
Bae
I. After incubation at 65o C for 15 min to inactivate
Bae
I, the reaction mixture was divided into two aliquots: one was mixed with stop
solution immediately (lane -); the other was treated with Klenow fragment at room temperature for 10
min and then mixed with stop solution (lane +). The two
Bae
I cleavage reaction products were loaded on an 8% denatured polyacrylamide gel
along with standard G, C, A and T ladders.
The cleavage patterns of all five
Bcg
I-like enzymes are shown in Figure
3
. They all share the following properties: (i) they all produce 3' overhangs; (ii) their cleavage sites (indicated by arrows) are all located symmetrically; (iii) they all contain a pair of adenines
(in bold type) which is located symmetrically relative to the center of the
double-stranded duplex that is being cleaved. Indeed the pair of adenines in the
Bcg
I recognition sequence is the target site for methylation (
3
,
4
). Thus, it is reasonable to speculate that the symmetrically located adenines
are the possible methylation target sites for all
Bcg
I-like enzymes, based on the presence and location of the two common
adenines.
Figure 3
.
All five
Bcg
I-like enzymes contain a pair of adenines (in bold type) which are located
symmetrically relative to the center of the double-stranded duplex that is being cleaved. The cleavage sites are indicated by
arrows.
To test this hypothesis, the target base of
Bae
I methylation was determined by
in vitro
methylation of a DNA duplex containing a
Bae
I site with this enzyme and [
3
H]AdoMet. The duplex DNA was degraded to mononucleotides, separated by thin-layer chromatography and visualized by unlabelled methyldeoxynucleotides which were
added to the reaction (
5
). The m
6
A, m
5
C and m
4
C spots were counted for radioactivity, >95% of the radioactivity co-migrated with m
6
A (Table
1
), indicating that
N
6-methyladenine is the sole product of
Bae
I methylation.
Table 1
.
Identification of the nucleosides methylated by
Bae
I
Solvent G
Solvent D
m
6
A
m
5
C + m
4
C
m
6
A + m
4
C
m
5
C
3H Incorporation
a
185
28
192
24
Purified
Bae
I enzyme complex was incubated with [
3
H]AdoMet and an unmodified duplex containing a
Bae
I site (Fig. 4). Following methylation, the duplex was digested to mononucleosides, mixed with unlabelled methyldeoxynucleosides,
and separated by thin-layer chromatography. Two different solvents were used to distinguish all
possible methylated bases. Solvent G is 66:33:1 isobutyric acid:water:ammonium
hydroxide (v/v); solvent D is 80:20 ethanol:water (v/v). The methylated deoxynucleosides are m
4
C:
N
4-methylcytosine; m
5
C: 5-methylcytosine; m
6
A:
N
6-methyladenine.
a
Counts per minute. Raw data; background counts of 24 c.p.m. have not been
subtracted.
There are three adenines within the
Bae
I recognition sequence of the duplex used in this assay: two on the top strand
(A21 and A29) and one on the bottom strand (A20) (Fig.
4
). To distinguish which adenines are methylated by
Bae
I, three more oligonucleotides were made (in addition to the two unmodified
oligonucleotides), containing a
N
6-methyladenine at each potential position, so that different combinations
of hemi-methylated duplex substrates can be produced as shown in Figure
4
.
Figure 4
.
The methylation and cleavage activities of
Bae
I enzyme on various duplex DNA substrates. Complementary single-stranded oligodeoxynucleotides (CCGCCAGGCAAAGGCCGTTA
AC
GTATC
GATAC
GT
o C for 3 min and then slowly cooling to 25o C. The DNA duplexes are presented schematically with only the
Bae
I recognition sequence being shown in the figure. The numbers indicate the
positions of underlined adenines from 5'-ends. Y, pyrimidine; R, purine. The three adenines within the
Bae
I recognition sequence were methylated each at a time in some duplexes, as
indicated by `-CH
3
'. DNA methylation activities on different substrates were determined by
incubating annealed duplex with
Bae
I is the presence of [
3
H]AdoMet, 10 mM Tris-HCl (pH 8.0), 5 mM Na
2
EDTA and 1 mM DTT. Acid-insoluble radioactivity was quantified by liquid scintillation counting.
All DNA duplexes were incubated with
Bae
I in separate reactions in the presence of Mg
2+
. The digestion products were visualized on 4% agarose gel.
Bcg
I and its closest relative, the type I methyltransferases, are maintenance
methyltransferases which highly prefer hemi-methylated DNA substrate for methylation (unpublished observation on
Bcg
I;
8
). Indeed
Bae
I shares this characteristic of maintenance methylases based on the following
observations. First, the
Bae
I was only able to methylate 1.1% of the potential sites on the unmodified
substrate which was sensitive to
Bae
I endonuclease in the presence of Mg
2+
(Fig.
4
, line 1). Second, the methylation of adenine 20 on the bottom strand of the
hemi-methylated duplex not only increases
Bae
I methylation activity by 30-fold, but also makes it resistant to the
Bae
I endonuclease digestion (line 3). The methylations of the two adenines on the
top strand have different effects: methylation on adenine 21 results a hemi-methylated duplex which is a preferred substrate for
Bae
I methylation (13-fold increase) and is insensitive to
Bae
I digestion (line 4); while methylation of adenine 29 on top strand generates a
false hemi-methylated substrate which behaves just like the unmodified one (line 5
versus line 1), indicating adenine 21 on the top strand is the appropriate
methylation target site not adenine 29. Finally, the lack of
Bae
I methylation activity on the DNA substrate with methylated adenine 21 on the
top strand and methylated adenine 20 on the bottom shows that
Bae
I cannot transfer any methyl group to the substrate from AdoMet when these two
proposed target sites are fully methylated. These results support our
speculation: the symmetrically located adenines existing in all
Bcg
I-like enzymes are the possible target bases for DNA methylation.
ACKNOWLEDGEMENTS
The authors would like to thank Drs Richard Roberts, Ira Schildkraut, Charles
Cantor and Cassandra Smith for critical reading of this manuscript, and Dr Peter Barrett for communicating his unpublished observation regarding the
Bae
I cleavage site.
REFERENCES
1 Degtyarev ,S.K. , Rechkunova,N.I., Zernov,U.P., Dedkov,V.S., Chizikov,V.E., Calligan,M.V., Williams,R. and Murray,E. (1993 ) Gene 131 93 -95.MEDLINE Abstract
2 Friedman ,S. (1986 ) Nucleic Acids Res 14
3 Kong ,H. , Morgan,R.D., Maunus,R.E. and Schildkraut,I. (1993 ) Nucleic Acids Res. 21
4 Kong ,H. , Roemer,S.E., Waite-Rees,P.A., Benner,J.S., Wilson,G.G. and Nwankwo,D.O. (1994 ) J. Biol. Chem. 269 683 -690.MEDLINE Abstract
5 Landry ,D. , Looney,M.C., Feehery,G.R., Slatko,B.E., Jack,W.E., Schildkraut,I. and Wilson,G.G. (1989 ) Gene 77
6 Roberts ,R.J. and Macelis,D. (1996 ) Nucleic Acids Res 23 223 -235.
7 Sanger ,F. , Nicklen,S. and Coulson,A.R. (1977 ) Proc. Natl. Acad. Sci USA 74 5463 -5467.MEDLINE Abstract
8 Suri ,B. , Nagaraja,V. and Bickle,T.A. (1984 ) Curr. Top. Microbiol. Immunol. 108
9 Vitor ,J.M.B. and Morgan,R.D. (1995 ) Gene 157 109 -110.
10 Yuan ,R. (1981 ) Annu. Rev. Biochem 50 285 -315.MEDLINE Abstract
Return
*
To whom correspondence should be addressed
