On the Nucleotide Sequence Recognized by a Eukaryotic Site-specific Endonuclease, Endo.Sce1 from Yeast*

Endo.Sce1 which is isolated from cells of Saccharo- myces cerevisiae is a eukaryotic site-specific endonuclease active on double-stranded DNA. At each cleavage site, Endo.Sce1 cuts only a defined phosphodiester bond in each strand of the double helix. We compared nucleotide sequences around five cleavage sites for Endo.Sce1 using a computer. We could not find any common specific sequence consisting of five base pairs or more among them. However, we found a 26-base pair consensus sequence which included 15 conserved nucleotides, allowing any of the five sequences to include a few nucleotides deviated from the consensus sequence. The consensus sequence is 5‘-CAn*PYnn-AniiCYYGTTnnnPnYnnYA-3’, where P, Y, n, and * denote purine, pyrimidine, any nucleotide, and the center of the cleavage site, respectively. The numbers of sites at which the consensus sequence appears in pBR322 DNA, dX174 replicative form DNA, fd repli- cative form DNA, or SV40 DNA are close to those of the cleavage sites for Endo.Sce1. We found that a 33- base pair fragment was efficiently cut at the defined phosphodiester bonds by Endo.Sce1. This 33-base pair fragment included 25 base pairs out of the 26-base pair consensus sequence. The fragments in which a part of

recombination in eukaryotes (Holliday, 1974;Radding, 1978). The frequencies of gene conversion within a gene always indicate a gradient; high at one end and low at another end (Hastings and Whitehouse, 1964). This polarity in the frequency suggests that the formation of the heteroduplex joint starts at a specific site outside of the gene. This site-specific initiation is easily explained if the cells have an endonuclease (site-specific endonuclease) which cleaves DNA at initiation sites (Angel et al., 1970;Catchside and Angel, 1974;Holliday, 1974;Radding, 1978). We had looked for eukaryotic sitespecific endonucleases, since no site-specific endonucleases had been known in eukaryotes in those days. We have found eukaryotic site-specific endonucleases in several strains of yeasts (Watabe et al., 1981) and purified one (Endo.Sce1) of them to apparent homogeneity from Saccharomyces cereuisine (Watabe et al., 1983;Watabe et al., 1984). Like prokaryotic type I1 restriction endonucleases, Endo.Sce1 cuts doublestranded DNA at strictly defined sites.
In the case of most type I1 restriction endonucleases, a palindromic specific sequence consisting of 4 base pairs or more is found at or near the cleavage sites and is specific to each endonuclease (see Roberts, 1982 for review). The nucleotide sequences at the cleavage sites for some type I1 restriction endonucleases (HphI, HgaI, MboII, and TthlllII) and type 111 restriction endonucleases (EcoPI,EcoP15,and HinfI 11) are heterogeneous. In these cases, a specific asymmetric sequence consisting of 5 to 6 base pairs was found for each endonuclease some base pairs away from the cleavage site. The number of base pairs between the specific sequence and the cleavage site is usually fixed for each enzyme (Kleid et al., 1976;Brown and Smith, 1977;Brown et al., 1980;Shinomiya et al., 1980;Haberman, 1974;Reiser and Yuan, 1977;Kauc and Piekarowicz, 1978). Type I restriction endonuclease recognizes the presence or absence of modification (methylation) in a specific sequence and, if the sequence is not modified, cuts double-stranded DNA at random sites in the presence of ATP and S-adenosylmethionine. Therefore, restriction endonuclease strictly recognizes a relatively short specific nucleotide sequence for the cleavage of DNA. On the other hand, the regulation of gene expression involves specific interaction of regulatory protein and DNA at the regulatory region, eg. promoter sites and RNA polymerases, or operator sites and repressor proteins (see Rosenberg andCourt, 1979 andLittle andMount, 1982). A consensus sequence for each protein was found in regulatory regions of various genes (see Rosenberg andCourt, 1979 andLittle andMount, 1982 for review). Unlike the case of restriction endonucleases, many of the sequences have some diversity from the standard consensus sequences and the diversity seems to be important in their regulatory functions.
We have analyzed nucleotide sequences around three cleavage sites for Endo.Sce1, one site in pBR322 DNA, and two sites in phage #X174 RF' DNA, and found that these nucleotide sequences were apparently heterogeneous (Watabe et al., 1983). Then, we further analyzed the nucleotide sequences around two cleavage sites in fd RF DNA. We compared these two sequences and the three sequences previously analyzed and found a consensus sequence among these five sequences.

MATERIALS AND METHODS
DNAs and Enzymes-RF DNAs of phages 6x174 and fd, and plasmid pBR322 DNA were prepared as described or cited previously (Watabe et al., 1981;Watabe et al., 1983).
Bacterial alkaline phosphatase and polynucleotide kinase were purchased from Bethesda Research Laboratories and Takara Shuzo Co. (Kyoto), respectively. Restriction endonucleases were from Bethesda Research Laboratories and New England Biolabs.
Endo.Sce1 was highly purified as follows. Cells harvested from a late log phase culture in a medium containing Polypeptone, yeast extract, and glucose were disrupted by French Press, and the cell-free extracts obtained were fractionated by polymin-P, the first and second phosphocellulose column chromatographies, and successive column chromatographies on DEAE-cellulose, double-stranded DNAcellulose, and heparin-Sepharose. The details of the purification of Endo.Sce1 was described in Watabe et al. (1984).
Treatment of DNA with Endo.SceI-The standard reaction mixture (40 pl) consisted of 50 mM Tris-HC1 (pH 7.5), 10 mM MgC12, 50 mM KCl, 5 mM 2-mercaptoethanol, 14 p~ (in nucleotides) DNA, and 1 to 3 units of Endo.Sce1, unless otherwise stated. After incubation at 37 "C for 60 min, the reaction was terminated by chilling in an icewater bath followed by the addition of 6 pl of a "stop-mixture" which consisted of 0.1 M EDTA, 60% sucrose, and 2% sodium dodecyl sulfate.
One unit of Endo.Sce1 is defined as the minimum amount of Endo.Sce1 required for digestion of Bacillus phage M2 DNA to produce the fourth largest SceI fragment' (Watabe et al., 1984).
Determination of Nucleotide Sequences around the Cleavage Sites for Endo.SceI-Fragments that contained only one cleavage site for Endo.SceI were prepared by cleaving double-stranded DNA with appropriate combinations of restriction endonucleases and purifying the products by gel electrophoresis. These restriction fragments' were treated with alkaline phosphatase and both 5"termini of the fragments were labeled with "P using [ Y -~~P J A T P and polynucleotide kinase 4Maxam and Gilbert, 1980). These fragments were treated with appropriate restriction endonuclease which cut the fragments at one site. The products were separated by gel electrophoresis to obtain the DNA fragments which contained one cleavage site for Endo.Sce1 and which were labeled with "P at only one of the 5"termini. Then, the fragments were cleaved either with Endo.Sce1 followed by heating at 90 "C for 1 min or by the base-specific chemical method of Maxam and Gilbert (Maxam and Gilbert, 1980). The treated fragments were subjected to electrophoresis through sequencing gel under the denaturing condition (Maxam and Gilbert, 1980). Autoradiographs of the gel were made using Fuji x-ray film at -80 "C.
Computer Analysis of Nucleotide Sequences-Nucleotide sequences were analyzed by a FACOM M380 system using a program package NASAR composed by one of us (T. K.).

Analysis of the Nucleotide Sequences around the Cleavage
Sites for Endo.SceI in fd RF DNA-We analyzed the nucleotide sequences around two cleavage sites in fd RF DNA. Double-stranded restriction fragments containing site A or site B were labeled with 32P at the 5"terminus of either the plus or minus strand. These labeled fragments were cleaved with Endo.Sce1 and analyzed by gel electrophoresis under the denaturing condition (Maxam and Gilbert, 1980). The treatment with Endo.Sce1 gave only one new band on the gel in

" "
The abbreviation used is: RF, replicative form. SceI fragments and HapII-TaqI fragments denote the fragments produced by cleavage with Endo.Sce1 and those with restriction endonucleases Hap11 and TaqI, respectively. The DdeI-TaqI fragment, for example, is defined similarly. Restriction fragments indicate double-stranded DNA fragments prepared by digestion with type I1 restriction endonucleases. all four cases tested; the plus strand of ThaI-AluI fragment' including site A (Fig. lA), a minus strand of the same ThaI-AluI fragment (Fig. lB), the plus strand of HapII-TaqI fragment including site B (Fig. IC), and a minus strand of DdeI-TaqI fragment including site B (Fig. 1D). These results clearly indicate that Endo.Sce1 cut only one defined phosphodiester bond in each strand at each cleavage site. The sequences around sites A and B in fd RF DNA are summarized in Fig.  2 4 . Like the cases of pBR322 DNA and 4x174 RF DNA (Watabe et al., 1983;see Figs. 2, B and C), Endo.Sce1 produced fragments with cohesive ends consisting of four nucleotides extending at the 3"termini.
To determine the phosphodiester bonds cleaved by EndoSceI, we compared the DNA fragments produced by Endo.Sce1 with those produced by the base-specific chemical methods of Maxam and Gilbert on sequencing gel electrophoresis in this study and the previous study (Watabe et al., 1983). The fragments produced by Endo.Sce1 have 3'-hydroxyl termini (Watabe et at., 1981;Watabe et al., 1983), while the fragments produced by the base-specific chemical method have 3'-phosphoryl termini (Maxam and Gilbert, 1980). This difference would cause the erroneous determination. However, we had confirmed that the simple comparison of the fragments produced by Endo.Sce1 with those obtained by the chemical method on the gel electrophoresis is reliable enough to determine the cleaved phosphodiester bonds by EndoSceI in the sequence; i.e. we had compared the results obtained by the above method with the direct analysis of 5"terminal nucleotides of SceI fragments by labeling with 32P using polynucleotide kinase when we had analyzed the sequences around pBR322 site and sites A and B in #X174 RF DNA. In all cases, we obtained the compatible results by these two methods which were based on different principles (Watabe et al., 1984).324 DNA Region Required for Recognition and Cleavage by Endo.SceZ-To deterpine the minimum region required for recognition and cleavage by Endo.Sce1, we tested DNA fragments with various chain lengths for the susceptibility to Endo.Sce1. First, we prepared a set of restriction fragments including the cleavage site for Endo.Sce1 from pBR322 DNA. Then, we treated these fragments with Endo.Sce1 and analyzed the products by gel electrophoresis under denaturing conditions. Endo.Sce1 cut a 33-base pair fragment (BstNI-RsaI fragment; Fig In all three cases, the fragments were cut at the same site located in the previous study ( Fig. 2B; see Watabe et al., 1983). The fragments used in the above experiments were labeled only at the 5"terminus of the minus strand (at the RsaI site), and the fragments treated with Endo.Sce1 were analyzed under denaturing conditions and detected by this radioactivity. Therefore, it seemed possible that Endo.Sce1 cut only the minus strand. Thus, we prepared the same 33base pair fragment (BstNI-RsaI fragment), labeled it only at the 5"terminus of the plus strand (at the BstNI site), and repeated the experiment. As shown in Fig. 3B, the plus strand T. Shibata, H. Watabe, T. Iino, and T. Ando, unpublished observations. method, the 5"termini of SceI fragments of 4x174 RF DNA were 'By the comparison with fragments produced by the chemical determined to be thymidine for the plus strand at site A and adenosine for the minus strand at site B (Watabe et al., 1983;see Fig. 2). By the direct analysis including labeling 5"terminal nucleotides, nucleotides at 5"termini of SceI fragments were thymidine (97% of all labeled nucleotides) for the plus strand at site A and adenosine (85%) for the minus strand at site B.

A Y P GScel
".

5'
FIG. 1. Analysis of the nucleotide sequence around the cleavage sites for Endo.Sce1 in fd RF DNA. A and B, analysis of nucleotide sequence around site A: A ThI-AluI fragment (positions 5912 to 6109), in which the 5"terminal nucleotide of the plus strand was labeled (at the ThaI site), was cleaved either with Endo.Sce1 or by the base-specific chemical method of Maxam and Gilbert. The treated fragments were subjected to electrophoresis through 8% sequencing gel after denaturation, and an autoradiograph of the gel was taken ( A ) . A ThaI-AluI fragment (positions 5912 to 6109), in which the 5"terminal nucleotide of the minus strand was labeled (at the AluI site), was treated as described above ( B ) . C and D, analysis of nucleotide sequence around site B: A HapII-Tag1 fragment (positions 2397 to 2530), in which the 5"terminal nucleotide of the plus strand was labeled (at the HapII site), was treated as described above (C). A DdeI-Tog1 fragment (positions 2364 to 2530), in which the 5'terminal nucleotide of the minus strand was labeled (at the Tag1 site), was treated as describe above (D). Y cleaved a t cytidine (C) or thymidine (7') (or weakly at G); P cleaved mainly a t adenosine ( A ) or guanosine (G); C cleaved a t cytidine; G cleaved a t guanosine; SceI cleaved with EndoSceI; c and 1, phosphodiester bond cleaved by EndoSceI. The sequences shown in A, B, C, and D (from the bottom of the prints) were read as 5"YYYYYGAYYAYYAAYYGGGGXAYAYAYJ,GAYYGA-3' ( a ) 5"GGYAATCGYAAAACTAGCATGYYAATGYYAATCATATJGYAYYYYGG-3' (b)

( C )
and 5"GCCTTYAGYGTCAGACJ,TGTAGCGCGYYYYYA-3' ( 4 respectively. In these sequences, P, Y, and X denote purine nucleotide, pyrimidine nucleotide, and a nucleotide that could not be identified, respectively. The sequences a, b, c, and d appear only between position 5982 and position 6014, between position 6035 and position 5996, between position 2465 and position 2486, and between position 2490 and position 2460, respectively, according to sequence analysis by a computer (see Fig. 2 4 ) . Therefore, we assigned the sequence as indicated.
of this fragment was also cut by Endo.Sce1 at the defined phosphodiester bond described previously ( Fig. 2B; see Watabe et al., 1983). These results indicate that the biochemical features sufficient for recognition and cleavage by Endo.Sce1 reside in the 33-base pair fragment.
We made similar experiments on the cleavage sites for Endo.Sce1 in 4x174 RF DNA and fd RF DNA, and the results are summarized in the lower half of Fig. 6. They indicate that a region including more than 10 base pairs on both sides of the cleavage sites is required to recognize and/or cleave double-stranded DNA by Endo.Sce1.
A Consensus Sequence around the Cleavage Sites for Endo.SceI-Then, we compared nucleotide sequences around five cleavage sites; two in fd RF DNA, one in pBR322 DNA, and two in 4x174 RF DNA using a computer. Since Endo.SceI cuts pBR322 DNA consisting of 4362 base pairs (Sutcliffe, 1979) a t one site, 4x174 RF DNA consisting of 5386 base pairs (Sanger et al., 1978) a t 2 sites, and fd RF DNA consisting of 6408 base pairs (Beck et al., 1978) a t 3 sites, and since Endo.SceI does not cut SV40 DNA consisting of 5226 base pairs (Fiers et al., 1978Reddy et al., 1978Buchman et al., 1980), the sites recognized by Endo.Sce1 were statistically expected to have a common sequence of 5 to 7 base pairs, if Endo.Sce1 recognized a specific nucleotide sequence in the same manner as restriction endonucleases. Unlike the case of type I1 or type 111 restriction endonuclease, there are no obvious specific nucleotide sequences consisting of five nucleotides or more at or near these five cleavage sites for Endo.Sce1 (Fig. 2). The possibility that the preparation of Endo.Sce1 contains five species of site-specific endonucleases is most unlikely, as discussed previously (Watabe et al., 1983;Watabe et al., 1984).
Since Endo.Sce1 exhibits strict site specificity in cleavage, it should recognize some common features in regions around the cleavage sites. Considering the finding described in the preceding section, we compared the sequences of the 50-base pair regions including a cleavage sites at the center, i.e. -25 to +25 region where position 0 is the center of cleavage site. Since each DNA consists of two antiparallel strands, we need to try two cases in order to compare the sequences of two  (1978). B, the whole nucleotide sequence of the smallest fragment which is cleaved by Endo.Sce1. This sequence was based on the data shown in Fig. 3. Numbers indicate the position of nucleotide in pBR322 DNA determined by Sutcliffe (1979).

CCGTTGAGTTCGATAATGGTGATATGTAT GTTGACGGCCATAAGGCTGCTTCTGACGTTCG CGCAACTCAAGCTATTACCACTATA CATACAACTGCCGGTATTCCGACGAAGACTGCAAGC
DNA molecules, i.e. (i) the plus strand of one DNA and the plus strand of the other, and (ii) the plus strand of one DNA and the minus strand of the other. Therefore, when we compare N species of DNA, we need to make 2N" combinations ("strand combinations") of the plus or minus strands. On the other hand, the sequence around the five cleavage sites for Endo.Sce1 did not share a common symmetrical structure. Therefore, if the sequences around the cleavage sites share a consensus sequence, a particular strand combination among all possible strand combinations would give the maximum fitting with respect to the sequence. We took the plus strand or the minus strand from each of the five 50-base pair regions and aligned them with respect to the center of the cleavage site and counted the number of the positions (conserved positions) at which all five sequences shared a nucleotide (A, T, G, or C). We tested all 16 (= 24) possible strand combinations, but we could not find any significant difference in the fitting among them. Then, we picked out four 50-base pair regions to make five possible "site combinations." In each site combination, we made eight (= 23) possible strand combinations, and for each strand combination we counted the number of the conserved positions as described above. As shown in Fig. 4, in any of the five site combinations, the maximum number of the conserved positions (i.e. the maximum fitting) always was obtained when we made strand combinations from the plus strands of the pBR322 site and 4x174 site A, and the minus strands of 4x174 site B, fd site A, and fd site B.
To see whether or not the maximum fitting found in Fig. 4 related to recognition of cleavage sites, we examined the extended region of 500 base pairs, using the plus strands of the pBR322 site and 6x174 site A, and the minus strands of 4x174 site B, fd site A, and fd site B. We aligned these five sequences with respect to the center of the cleavage site, and counted the number of positions at which all five sequences shared a nucleotide, or purine-or pyrimidine nucleotide in each of the 20-base pair subregions. As shown in Fig. 5, the position at which all five sequences share a nucleotide (A, T, G, or C) appears only in the -10 to +30 region, and the number of positions at which all five sequences share a purine nucleotide (P) or pyrimidine nucleotide (Y) is significantly larger in the same -10 to +30 region than any other region. These results strongly suggest the existence of a consensus sequence among the five sequences near the cleavage sites for Endo.Sce1 and that the sequence is asymmetric with respect to the cleavage site.
These results are also consistent with the finding that Endo.Sce1 cuts efficiently a 33-base pair fragment which covered -11 to +22 region. Therefore, we analyzed the -11 to +22 region and flanking regions of a few nucleotides of the same combination of strands including each of the five cleavage sites; i.e. the plus strands from the pBR322 site and 4x174 site A, and the minus strands from 4x174 site B, fd site A, and fd site B. We picked out the nucleotides at the positions where at least four strands shared a nucleotide (A, T, G, or C), or a purine nucleotide (P) or pyrimidine nucleotide (Y) and obtained a consensus sequence,

5'-CAn*PYnnAnnCYYGTTnnnPnYnnYA-3',
where * and n indicate the center of the cleavage site and any nucleotide, respectively (Fig. 6). At position -11, four out of the five strands have purine (Fig. 6). HDwever, the species of nucleotide at -11 seems not to be recognized by Endo.Sce1, since the position -11 is separated by the nonconserved region of seven base pairs from a core conserved region in the consensus sequence. Therefore, we ignored a purine at -11. Allowing any of the five strands to have at most two nucleotides deviating from a specific sequence, we consider the following sequences as candidates of the specific sequence recognized by Endo.Sce1:

5'-CAn*PYnnAnnCYYPTTnnnPnYnnYA-3' (Sequence 2)
The frequencies with which these sequences appear in pBR322 DNA, 4x174 RF DNA, fd RF DNA, and SV40 DNA are close to those of the cleavage sites (Table I). As suggested from the analysis shown in Fig. 5, this consensus sequence is asymmetric, and does not include palindrome or inverted repeats. cleaved with EndoSceI or by the base-specific chemical method of Maxam and Gilbert. C, a BstNI-RsaI fragment (positions 132 to 165), in which the 5"terminal nucleotide of the minus strand was labeled (at the RsaI site), treated as in B. The treated fragments were denatured and subjected to electrophoresis through sequencing gels (8% in A, and 20% in B and C ) , and autoradiographs of the gels were taken. Y cleaved a t cytidine ( C ) or thymidine (T); P cleaved at adenosine ( A ) or guanosine (C); C cleaved a t cytidine; G cleaved a t guanosine; SceI cleaved with Endo.Sce1.

DISCUSSION
By the comparison of the primary sequence around five cleavage sites for Endo.Sce1, we found a consensus sequence of 26 base pairs ( Fig. 6 and Table I), including 15 conserved nucleotides and covering the -3 to +23 region around the cleavage sites. Twenty five base pairs of this 26-base pair sequence are included in a 33-base pair fragment (BstNI-RsaI fragment of pBR322 site) which is efficiently cut by EndoSceI (Figs. 3 and 6). As expected from the hypothesis that this consensus sequence is recognized by Endo.Sce1, DNA fragments missing the right arm from position +11 or +10 (Hap11 fragment of fd site A, HhaI fragment of fd site B) were not cut by Endo.Sce1 (Fig. 6). A fragment missing the left arm from position -12 (BstNI-RsaI fragment of the pBR322 site) was cut efficiently by Endo.Sce1, but a fragment missing the left arm from position -11 was not cut by the endonuclease (Fig. 6). This indicates that the region of -4 to -11 is required for the cleavage. The species of nucleotide a t -11 seems not to be recognized by EndoSceI, as discussed in the previous section.
It should be noted that the fact that the 33-base pair fragment, which covers position -11 to position +22, is cut by Endo.SceI does not rule out the possibility that the sequence outside of the -11 to +22 region is also involved in the recognition, because the presence of flanking sequences might help the recognition by preventing the binding and/or cleavage unless the sequences fit the endonuclease. This seems to explain partly the discrepancy between the number of recognition sites calculated from a proposed recognition sequence and the actual number of cleavage sites (see Table I).
Any of the five strands around cleavage sites for Endo.Sce1 has one to three nucleotides which do not fit the consensus 0 X A k + + t t t + t \ *+t+"" +++t"--+t++----  indicate purine nucleotide, pyrimidine nucleotide, and the center of the cleavage site, respectively, and n indicates any nucleotide. The number of sites in DNA was examined with a computer using the following principles: (i) at most 2 nucleotides were allowed which deviated from the sequence indicated, and (ii) T and G were allowed a t 5'-and 3'-termini, respectively, as one of the nonmatching nucleotides.

fd E + t --t t --++--tt--t t --t t --
suggests that the site specificity of Endo.Sce1 can be controlled by conditions, such as cations (Watabe et al., 1984).' These characteristics will be important if Endo.Sce1 plays a regulatory role in a cellular function, such as the initiation of recombination.
Recently, two new site-specific endonucleases, YZ-Endo (HO-Endonuclease) and EndoSceII, were found from cellfree extracts of S. cereuisiae (Kostriken et al., 1983). YZ-Endo is shown to play a role in the initiation of mating type switching, a type of genetic recombination, in homothallic strains of yeasts. The cellular function of Endo.SceI1 is not known. Both YZ-Endo and Endo.SceI1 act on double-stranded DNA and make double-strand scission. Since YZ-Endo did not cut pBR322 DNA, and Endo.SceI1 cut the DNA at one site near the AuaI site, YZ-Endo and Endo.SceI1 appear to have site specificities different from that of Endo.Sce1 which cut pBR322 DNA a t one site near the Hind111 site. It would be worthwhile to note that the structure of cohesive ends produced by either YZ-Endo and Endo.SceI1 is the same as that by Endo.Sce1; i.e. the cohesive ends formed by any of these endonucleases consist of four nucleotides extending at 3'-termini ( Fig. 2; see Watabe et al., 1983 andKostriken et al., 1983).
Moreover. when we comDared the sequence around the sequence (Fig. 6). Therefore, Endo.Sce1 seems to recognize a primary sequence allowing some limit of deviation from a standard specific sequence. This mode of recognition is different from that of restriction endonucleases but rather resembles the mode of recognition of the target sites by regulatory proteins (see "Introduction," and Rosenberg andCourt, 1979 andLittle andMount, 1982 for review). It is likely that the extent of deviation in the primary sequence of a site from a standard specific sequence determines the susceptibility of the site to cleavage by Endo.Sce1. This prediction is supported by the observation that DNA of phages X, 4105, and M2, all of which had several cleavage sites for Endo.Sce1, were never cleavage site for YZ-Endo 0; Endo.SceI1 with the consensus sequence for Endo.Sce1, we found some extent of homology.
In the plus strand of the cleavage site in pBR322 DNA for Endo.SceI1, the +1 to +11 region exactly fits the +1 to +11 region of a consensus sequence for Endo.Sce1, while the minus strand has no such homology (Fig. 6) gagcttaat CA a G& t g A t g CTCGTT a t g G t T t c Cg + 6x174 s i t e ggccgtcaa CA t AC a t A t c 5CCnTT a t c G a & c t CA -6x174 s i t e catgtcaat CA t AT g t A cc CCgGTT g a t A a T ca &A f d s i t e A c t t t a g c g t CA g AC t g 4 ag CgCGTT t t c A t C gg CA f d s i t e B

10505
A B CA n*PY M A nn CYYGTT nnn P n Y nn YA Consensus sequence gaattggag C c a AT ca A tt C T T G Q gag A a C t g TC gcaagaatt g A t 48 gc 4 cc g T S T tgg A g T gg Tg a g c t t t c c g CA a g t A aa aTTLT& taa A c C c t @ YZ-hdo a a a a t t t t a C& g 4 T gc g ga -T& aaa c t 5 aa e 1) a t MATa-YZ

5'BstNI)
Cut by Endo.= Upper case without underline in the sequence indicates a nucleotide which fits the consensus sequence. Upper case with underline indicates a nucleotide which does not fit the consensus sequence but belongs to the same group (purine or pyrimidine) as that in the consensus sequence. Lower case with underline indicates a nucleotide which deviates from the consensus sequence. P, Y, and n denote purine nucleotide, pyrimidine nucleotide, and any nucleotide, respectively. In the lower half, bars indicate the strands of restriction fragments used in the test for their susceptibility to the cleavage by Endo.Sce1. Name of restriction endonucleases at the termini of the bars indicates the endonuclease to form their termini. The center of the cleavage sites for Endo.Sce1 are as follows: between 143 and 144 for the pBR322 site, 1462 and 1463 for a @X174 site A, 4198 and 4197 for a @X174 site B, 6007 and 6006 for fd site A, and 2477 and 2476 for fd site B (Watabe et al., 1983;Figs. 1 and 2). The sequence around the cleavage site for Endo.SceI1 or YZ-Endo was determined by Kostriken et al. (1983). The center of the cleavage site in pBR322 DNA for Endo.SceI1 is between 1329 and 1330.
G are nucleotides fitting the consensus sequence, x is a nucleotide not fitting the consensus sequence, and = is a nucleotide not conserved in the consensus sequence. The minus strand of the -3 to +23 region of the site is 5'-Cn=*rT==n==rxnxTx===x=x==n(G)-3'.
This large difference in homology to the consensus sequence for Endo.Sce1 between the plus strand and the minus strand support the idea that the homology of the sequence around the cleavage site for YZ-Endo and the consensus sequence for Endo.Sce1 is not accidental. The similarity in the structure of cohesive ends formed by the cleavage, and the homology in the primary sequence around cleavage sites suggest the functional similarity and evolutional relationship of these three site-specific endonucleases, Endo.Sce1, Endo.SceI1, and YZ-Endo in S. cereuisiae.