Skip Navigation

This Article
Right arrow Abstract Freely available
Right arrow Print PDF (328K) Freely available
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (4)
Right arrowRequest Permissions
Right arrow Commercial Re-use Guidelines
for Open Access NAR Content
Google Scholar
Right arrow Articles by Rice, M. R.
Right arrow Articles by Blumenthal, R. M.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Rice, M. R.
Right arrow Articles by Blumenthal, R. M.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?

Nucleic Acids Research Pages 1032-1038  

Substrate recognition by the Pvu II endonuclease: binding and cleavage of CAG5mCTG sites
Introduction
Materials And Methods
   Bacterial transformations
   Synthetic DNA substrates
   Cleavage of synthetic oligonucleotides
   DNase I footprinting
   Analysis of the specificity of R·PvuII binding to CAG5mCTG
Results
   aluIM and pvuIIR cannot stably coexist in E.coli
   R·PvuII cleaves synthetic DNA that contains CAG5mCTG on both strands
   R·PvuII binds to CAG5mCTG sites in footprinting assays
   Competition assays reveal reduced affinity for CAG5mCTG compared with the unmethylated sequence
Discussion
   Cleavage of CAG5mCTG sites
   Region of DNA contacted by R·PvuII
   Binding of CAG5mCTG by R·PvuII
   The R·PvuII:DNA crystal structure is consistent with cleavage of CAG5mCTG sites
Acknowledgements
References

Substrate recognition by the Pvu II endonuclease: binding and cleavage of CAG<sup>5m</sup>CTG sites

Substrate recognition by the Pvu II endonuclease: binding and cleavage of CAG5mCTG sites

Manda R. Rice, Michelle D. Koons+ and Robert M. Blumenthal*

Department of Microbiology and Immunology, Medical College of Ohio, 3055 Arlington Avenue, Toledo, OH 43614-5806, USA

Received October 13, 1998; Revised November 9, 1998; Accepted December 15, 1998

ABSTRACT

The PvuII restriction endonuclease (R·PvuII) cleaves CAG[darr]CTG sequences as indicated, leaving blunt ends. Its cognate methyltransferase (M·PvuII) generates N4-methylcytosine, yielding CAGN4mCTG, though the mechanism by which this prevents cleavage by R·PvuII is unknown. The heterologous 5-methylcytosinemethylation CAG5mCTG has also been reported to prevent cleavage by R·PvuII and this has been used in some cloning methods. Since this heterologousmethylation occurs at the native methylated base, it can provide insights into the detection of DNAmethylation by R·PvuII. We found that the cloned gene for R·PvuII could not stably transform cells protected only by M·AluI (AG5mCT) and then determined that R·PvuII cleaves CAG5mCTG in vitro, even when both strands are methylated. DNase I footprint analysis and competition experiments reveal that R·PvuII binds to CAG5mCTG specifically, though with reduced affinity relative to the unmethylated sequence. These results provide biochemical support for the publishedstructures of R·PvuII complexed with DNA containing CAGCTG and CAG5-iodoCTG and support a model for how methylation interferes with DNA cleavage by this enzyme.

INTRODUCTION

The PvuII restriction-modification system (1) has been cloned and sequenced (2-5) and subjected to regulatory studies (5-7). This system is the first to have had structures determined for both the endonuclease (R·PvuII; 8-10) and the methyltransferase (M·PvuII; 11,12). While a number of mutants have been generated in the gene for R·PvuII (13), this enzyme is less well understood kinetically than its cognate methyltransferase (MTase; 14). M·PvuII acts on the exocyclic amino moiety of the internal cytosine in the sequence CAGCTG, generating N4mC (2,15), and this methylation prevents DNA cleavage in vivo by R·PvuII. One of the major questions about restriction endonucleases in general, and R·PvuII in particular, is how DNA methylation is recognized and how it inhibits the cleavage reaction.

In addition to being inhibited by the native methylation carried out by the cognate MTase, many restriction endonucleases are also inhibited by heterologous methylation. AluI MTase (M·AluI) modifies AGCT sites at the same cytosine as does the PvuII MTase (CAGCTG), though the methyl group is added to a ring carbon, generating 5mC (16). There are three published sources reporting that R·PvuII does not cut AluI-methylated DNA. In one (17), 2 µg of chromosomal DNA from Arthrobacter luteus, the host strain for the AluI system, was not digested by 4 U of R·PvuII in 3 h. In the second (15), no digestion conditions were given but it was reported that R·PvuII did not cleave plasmid DNA that had been methylated in vitro by M·AluI. Finally, a third study (18) reported that 8 U of R·PvuII failed to cleave an unspecified amount of AluI-methylated pBR322 DNA within 3 h.

The ability of heterologous methylation to prevent cleavage by restriction endonucleases has been exploited in a variety of ways. For example, methylation of the substrate DNA can be used to generate new endonuclease specificities, including specificities 12 bp long. These applications and a summary of methylations that prevent cleavage by various endonucleases, have been reviewed (19). Some of these applications, however, depend on complete protection from cleavage and this is not always provided by heterologous methylation.

We had intended to use M·AluI to protect DNA from R·PvuII cleavage, in order to select for mutants of R·PvuII with relaxed (NAGCTN) specificity. For this purpose we attempted to create a strain carrying both aluIM (the gene for M·AluI) and pvuIIR (the gene for R·PvuII), but not pvuIIM (the gene for M·PvuII). Because this strain proved to be genetically unstable and because others have found that heterologous 5mC in some cases slows but doesn’t prevent cleavage by various endonucleases (19,20), we decided to investigate the possibility that R·PvuII cleaves M·AluI-methylated DNA. In the context of studies on substrate recognition by R·PvuII we found that this enzyme protects a discrete region on each DNA strand from DNase I, whether or not 5mC is present at the internal cytosine and that fully methylated CAG5mCTG sites are cleaved at a significant rate.

MATERIALS AND METHODS

Bacterial transformations

The plasmids used in these experiments are described in Table 1. Escherichia coli strain JM107MA2 (2) with and without plasmids was grown in LB medium to a Klett reading of 60 (corresponding to an A600 nm of 0.25). The cells were made competent with polyethylene glycol as described (21), transformed with 100 ng of plasmid DNA and were plated onto LB agar containing carbenicillin (Cb, 120 µg/ml) or kanamycin (Km, 50 µg/ml).

Table 1. Plasmids used
Plasmid Vector Relevant phenotype Antibiotic resistance Source Reference
pAluM2.7Kma pUC8 M·AluI+ Km This laboratory (39)
pPvuRM3.4 pBR322 R·PvuII+ M·PvuII+ Cb This laboratory (2)
pPvuRM3.3b pBR322 R·PvuII+ M·PvuII- Cb This laboratory (40)
pPvuM1.9Kma pUC8 M·PvuII+ Km This laboratory (2)
pBluescript KSc (ColEI origin)   Cb Stratagene Inc.  
aThe antibiotic resistance conferred by this plasmid was changed from the original construct for purposes of dual selection. Note that aluIM is expressed from its native Arthrobacter promoter and not from any heterologous strong promoter.
bAn in-frame SpeI deletion inactivates pvuIIM.
cpBluescript has the same origin of replication and carries the same antibiotic resistance as pUC8, and was used as a positive control for transformation.

Synthetic DNA substrates

The sequences and structures of the synthetic oligonucleotides are shown in Figure 1. A PCR product containing 5mC was generated using d5mCTP (US Biochemical, Cleveland, OH) in PCR reactions amplified with Taq DNA polymerase (Gibco BRL, Grand Island, NY) and plasmid pBR322 as a template (pBR322 contains a single PvuII site). The primers used were 5[prime]-CCC CCT TAC ACG GAG GCA-3[prime] and 5[prime]-GGC TGC GCC CCG ACA CCC-3[prime]. The oligonucleotides Pvu40R, Pvu40M and Pvu40X were purchased from New England Biolabs (Beverly, MA). The purity of the 5mC phosphoramidite was reported to be >99% and all oligonucleotides were PAGE purified by the manufacturer.

Cleavage of synthetic oligonucleotides

Pvu40R and Pvu40M were radiolabeled separately using T4 poly-nucleotide kinase (Gibco BRL) and [[gamma]-32P]ATP (6000 Ci/mmol; Amersham, Arlington Heights, IL) and unincorporated nucleotides were removed using QiaQuick Spin Columns (Qiagen, Chatsworth, CA). The concentration of radiolabeled DNA was determined from the A260 nm. Equimolar amounts of upper and lower strands were annealed by heating to 90°C and slowly cooling to room temperature. The specific radioactivity was determined as previously described (22). Cleavage assays were performed using 10 nM duplex DNA (11.4 ng) and either 10 pM (0.075 U) R·PvuII dimer for the unmodified duplex or 500 pM R·PvuII (3.75 U) for the modified duplex in a volume of 50 µl, using buffer recommended by the manufacturer and a glycerol concentration of 5% (v/v). R·PvuII (50 000 U/ml) was diluted in Diluent A (New England Biolabs) immediately before use. Reaction components were preincubated at 37°C for 2 min before adding R·PvuII. Portions were withdrawn, added to stop/loading dye (New England Biolabs) and placed in a dry ice-ethanol bath. These samples were heated to 90°C for 3 min, loaded onto a 20% sequencing gel cast in glycerol-tolerant buffer (23) and electrophoresed at 70 W. Data was collected from dried gels with a Storm 840 Phosphorimager (Molecular Dynamics, Sunnyvale, CA). Rates of cleavage were determined from an inverse exponential fit.


Figure 1. Synthetic oligonucleotides used. Oligonucleotides used in the experiments were designed based on the single PvuII site in plasmid pBR322. The PvuII sites are underlined. The M represents 5mC (Pvu40M). In Pvu40X the PvuII site has been reversed (CAGCTG to GTCGAC).

DNase I footprinting

DNase I footprinting of R·PvuII:DNA complexes was performed with a 770 bp NdeI-AlwNI fragment from plasmid pSP70 (Promega, Madison, WI). This fragment, which contains a single PvuII site 90 bp from the NdeI site, was gel purified using the QiaQuick kit (Qiagen), radiolabeled on the 3[prime]-end by filling in the NdeI site with Klenow polymerase and [[alpha]-32P]dATP and separated from unincorporated label using QiaQuick spin columns (Qiagen). The radiolabeled DNA (160 pM) was mixed with purified R·PvuII (0-0.5 nM; a gift of X. Cheng) in a volume of 50 µl including 20 mM HEPES, 2 mM EDTA, 30 mM KCl, 10 mM CaCl2 and 5% glycerol, pH 7.5. Ca2+ is required for DNase I activity, but does not support cleavage by R·PvuII (unpublished observation) and no Mg2+ was added. The reactions were incubated at room temperature for 30 min, then 8.5 × 10-3 U of DNase I (Gibco BRL) was added. After 2 min the reactions were stopped by precipitating the DNA (ethanol, tRNA and saturated ammonium acetate) in a dry ice-ethanol bath. The DNA was pelleted, washed with 70% ethanol and resuspended in loading dye. The samples were electrophoresed as described above. Maxam-Gilbert sequencing reactions were performed using a kit (Sigma Chemical Co., St Louis, MO) according to the manufacturer’s recommendations.

Analysis of the specificity of R·PvuII binding to CAG5mCTG

Competition assays were performed to estimate the relative affinities of R·PvuII for unmodified and 5mC-modified oligonucleotides. The rate of cleavage of radiolabeled Pvu40R was measured as described above, in the presence of competitor DNAs ranging from 2- to 100-fold excess over the concentration of radiolabeled substrate. Competitor DNA was pre-mixed with the radiolabeled substrate DNA and reactions were initiated with R·PvuII. Competitor DNAs used included Pvu40R, Pvu40X and Pvu40M duplexes. Reactions were resolved on 20% sequencing gels as described above.

RESULTS

aluIM and pvuIIR cannot stably coexist in E.coli

If R·PvuII cannot cleave CAG5mCTG sites, then cells carrying aluIM should be able to stably carry pvuIIR in the absence of pvuIIM. However plasmids carrying the aluIM and pvuIIR genes could not be co-transformed (pAluM2.7Km × pPvuRM3.3; Table 2). Surprisingly, the co-transformation of a plasmid carrying aluIM was inefficient even in the presence of pvuIIM. The low transformation efficiency of pAluM2.7Km and pPvuRM3.4 (which produces M·PvuII as well as R·PvuII) may be due to the fact that DNA modified by M·AluI is not a substrate for M·PvuII (15) and this could lead to accumulation of CAGCTG sites methylated only by M·AluI.

Since the aluIM and pvuIIR genes were introduced together, however, perhaps M·AluI simply didn’t have enough time to accumulate and act on the cell DNA before active R·PvuII dimers accumulated. To test this possibility, JM107MA2 (pAluM2.7Km) was first established and then in a second step was transformed with pPvuRM3.3 with selection for KmRCbR colonies (Table 3). We found that when aluIM is introduced first, transformation with pvuIIR is possible (though the resulting colonies are unusually small). This is reminiscent of the cloning of certain restriction modification systems in which the methylase had to be cloned before the endonuclease (24,25).

These aluIM+pvuIIR+ co-transformants were tested for the status of the pvuIIR gene because the aluIM+pvuIIR+ colonies were much smaller than the pvuIIM+pvuIIR+ colonies and because we had previously observed pvuIIR::IS10 mutants in attempted pvuIIR+pvuIIM- transformations (5). Ten co-transformants of pvuIIR with each of the two MTase genes were quadruply passaged on plates selective for the pvuIIR plasmid. These colonies were then tested for R·PvuII activity by using a crude extract from overnight cultures to digest [lambda] phage DNA (data not shown). All 10 of the pvuIIM/pvuIIR colonies displayed R·PvuII activity. In contrast, only seven of the 10 aluIM/pvuIIR strains had detectable endonuclease activity. The presence of seven clones with R·PvuII activity indicates that M·AluI provides substantial protection against cleavage by R·PvuII, though this protection must be incomplete since there appears to be selection for pvuIIR mutants. It is worth noting that DNA from cells bearing the aluIM gene is fully resistant to digestion with R·AluI (26). Nevertheless, two possibilities remained open: that M·AluI fails to completely methylate the DNA in these cells or that even fully methylated CAG5mCTG sites are cleaved by R·PvuII.

R·PvuII cleaves synthetic DNA that contains CAG5mCTG on both strands

We demonstrated that DNA methylated in vitro with excess M·AluI is cleaved by R·PvuII (not shown). This cleavage was slow, relative to unmethylated DNA. As with the co-transformation experiments, one possible explanation for these results is that the in vitro modified DNA might have been incompletely methylated. Every AGCT site was at least hemimethylated, as the DNA was fully resistant to digestion with R·AluI, but hemimethylated sites containing 5mCAG5mCTG on one strand have been shown to be cleaved by R·PvuII (27).

In order to rule out the presence of unmodified or hemimethylated strands of DNA in the cleavage reactions, we took two approaches to synthesize DNA substrates in which the methylation status of the PvuII site was less ambiguous. In the first approach, DNA substrates were generated by PCR using d5mCTP in place of dCTP. This reaction generates substrates that are modified at all cytosines (5mCAG5mCTG) on both strands. The 300 bp PCR products were digested with R·PvuII and the products analyzed under both non-denaturing and denaturing electrophoreticconditions. Under the conditions used, no cleavage or even nicking of 5mCAG5mCTG was observed (not shown). However, this did not rule out cleavage of CAG5mCTG sites.

Table 2. Plasmids carrying pvuIIR and aluIM cannot be co-transformeda
Plasmid A Plasmid B
None pBlueScript pAluM2.7Km pPvuRM3.4 pPvuRM3.3
None 0b 6565 ± 230 3419 ± 231 1699 ± 297 0.3 ± 0.3
pAluM2.7Km 3419 ± 231 NDc   19.7 ± 0.3 0
aCo-transformations were performed as described in Materials and Methods. Escherichia coli strain JM107MA2 was co-transformed by 100 ng of each of the pairs of plasmids (A and B) as shown. Double transformants were selected on agar medium containing appropriate antibiotics. Colonies were counted after 16 h incubation at 37°C.
bResults shown are from triplicate samples. The number of transformants is expressed as the mean ± SE. Data are normalized for differences in the size of the plasmids. Negative controls were not performed in triplicate.
cND, not determined.

Table 3. pvuIIR can transform a strain already expressing aluIMa
Strain Plasmid
pBlueScript KS (10 ng) pPvuRM3.3 (10 ng) pPvuRM3.3 (100 ng) None
JM107MA2 7793 ± 565 0 0 0b
JM107MA2 (pPvu1.9Km) 4820 ± 730 1784 ± 362 4289 ± 85 0
JM107MA2 (pAluM2.7Km) 2103 ± 327 637 ± 64 2527 ± 584 0
aTransformations were performed as described in Materials and Methods. Strains were grown to mid logarithmic phase and then transformed with the plasmids shown. Transformants were selected on agar medium containing the appropriate antibiotics. Colonies were counted after 16 h growth at 37°C. Results shown are from triplicate samples. Number of transformants is expressed as the mean ± SE.
bNegative controls were not performed in triplicate.

The second approach made use of complementary linear oligonucleotides that were or were not synthesized with a d5mCTP phosphoramidite at the single position within the PvuII substrate sequence (Fig. 1). The products of overnight digestion were resolved on a non-denaturing acrylamide gel with and without sample preheating to reveal nicking activity. R·PvuII cleaved fully modified CAG5mCTG sites (Fig. 2). Note that small amounts of the double-stranded intermediate products CD and EF are present in the heat-treated lanes. Heating the products of digestion of the unmodified and modified oligonucleotides produced similar patterns. If nicking of the DNA had occurred, it must have occurred equally on both strands. Similar results were obtained with synthetic hairpin oligonucleotides (results not shown).


Figure 2. R·PvuII cleaves a synthetic oligonucleotide containing CAG5mCTG. Pvu40R and Pvu40M duplexes were radiolabeled and digested with R·PvuII. In order to detect the products of single-stranded cleavage (nicking), a portion of each reaction was denatured at 95°C prior to electrophoresis. The autoradiograph of the dried gel is shown in the upper panel. The lower panel shows the DNA substrate and the products (A-F) of R·PvuII cleavage.

In order to determine the relative rates of cleavage ofCAGCTG and CAG5mCTG sites, the linear synthetic oligo-nucleotides were digested as described in Materials and Methods (Fig. 3). Unmodified substrate was cleaved at an overall rate of 1.38 pmol/min/unit enzyme under these conditions. 5mC-modified substrate was cleaved at an overall rate of 9.2 × 10-3 pmol/min/unit enzyme, 150-fold slower than the unmodified substrate. As noted for the overnight digestions, the two strands of DNA were cleaved at indistinguishable rates.


Figure 3. Digestion of synthetic oligonucleotides by R·PvuII. Pvu40R contains CAGCTG (A) and Pvu40M contains CAG5mCTG (B). The line shown is an inverse exponential fit; initial rates of cleavage were determined from this equation. R values were [ge]0.93. The reaction products were resolved on a sequencing gel; data was collected via phosphorimaging.

R·PvuII binds to CAG5mCTG sites in footprinting assays

While R·PvuII does cleave CAG5mCTG sites, the rate of cleavage is low relative to unmodified substrates. One explanation is that enzyme binding to the methylated site might be impaired. In order to test this possibility, DNase I footprinting was performed with plasmid DNA that was either mock methylated or methylated in vitro by M·AluI and with either the upper or lower strand of DNA labeled. The data shown is for the lower strand (Fig. 4).


Figure 4. DNase I footprints of R·PvuII on M·AluI-modified or M·HhaI-modified substrates. DNase I footprinting was performed using a 775 bp plasmid fragment from pSP70 that had either been methylated using M·HhaI (which recognizes the irrelevant sequence GGCC) or M·AluI (AGCT) to modify the PvuII site to CAG5mCTG. Lane 1 shows a digest of the DNA with R·PvuII, to indicate the center of the PvuII site. Lane 2 shows the G ladder of Maxam-Gilbert sequencing reactions of the substrate (note: this lane is composited with a different exposure than the other lanes). Lanes 3-6 represent a range of R·PvuII concentrations from 0-0.25 nM, using the M·HhaI-modified DNA substrate. Lanes 7-10 show the same R·PvuII range using the M·AluI-modified substrate. The arrow indicates the R·PvuII-induced DNase I hypersensitive site.

The binding of R·PvuII to the DNA induced a very strong hypersensitivity to DNase I upstream of the PvuII site (Fig. 4, arrow). The hypersensitive site, regardless of the basis for its appearance, provided a sensitive quantitative marker for R·PvuII binding to the DNA. By this measure and by the appearance of the protected region, the mock-methylated DNA is bound only marginally better by R·PvuII than is the M·AluI-methylated DNA, suggesting that the presence of 5mC in the DNA does not greatly affect the equilibrium binding properties of R·PvuII under the conditions used. Certainly, the effect of 5mC on binding is much less than on cleavage.

Competition assays reveal reduced affinity for CAG5mCTG compared with the unmethylated sequence

Competitive cleavage assays were performed in order to measure the relative affinity of R·PvuII for unmodified and 5mC-modified oligonucleotide substrates under reaction conditions. Cleavage rates of radiolabeled unmodified substrates were measured in the presence of 10- to 40-fold excess unlabeled competitor oligo-nucleotides (unmodified specific DNA, DNA with a reversed PvuII site or with a CAG5mCTG PvuII site). The data shown is representative (Fig. 5). Under the conditions used, a 10-fold excess of unlabeled specific DNA made cleavage of the radiolabeled substrate almost undetectable. The specificity of the reaction was established by using a competitor oligonucleotide that lacks a PvuII site but is otherwise identical to the substrate oligonucleotide; >20-fold more of this reversed-site duplex than of the site-containing duplex was required to achieve half-maximal inhibition. The affinities of R·PvuII for DNAs containing a reversed site or a doubly methylated CAG5mCTG site appear to be comparable with one another in this assay. Thus under reaction conditions, reduced affinity of R·PvuII for CAG5mCTG sites appears to contribute substantially to the lower cleavage rate seen with the modified sites.


Figure 5. Effects of various competitor DNAs on substrate cleavage by R·PvuII. (A) Cleavage assays were performed in the presence of specific, non-specific (reversed site) and 5mC-modified competitor DNAs, as described in Materials and Methods. Each competitor DNA was tested at 100, 200 and 400 nM together with the 10 nM radiolabeled specific substrate. Representative primary digestion data are shown: [open circle], no competitor; [closed circle], 200 nM Pvu40M (5mC competitor). (B) The slopes from experiments such as those shown in (A) are plotted against the competitor concentration: [open circle], Pvu40M (5mC competitor); [closed circle], Pvu40X (reversed site competitor); [open square], Pvu40R (specific competitor).

DISCUSSION

Cleavage of CAG5mCTG sites

This work demonstrates that R·PvuII slowly cleaves AluI-methylated DNA. The contrast between this finding and other reports (15,17-19) may reflect our having used more sensitive assays, different levels of R·PvuII or perhaps to our having assayed cleavage of different PvuII sites (i.e. having different flanking sequences).

Kang et al. (18) reported that PvuII did not cleave AluI-methylated DNA; however, they did report single-strand cleavage and accumulation of nicked AluI-methylated DNA with HindIII and SstI endonucleases, which recognize AAGCTT and GAGCTC, respectively. The fact that other endonucleases exhibit this type of behavior on heterologously methylated substrates may point to common elements underlying recognition and cleavage among endonucleases recognizing the core sequence AGCT (28).

Region of DNA contacted by R·PvuII

Footprint analysis of R·PvuII has not previously been reported. The footprint appears to be 21 nt long (Fig. 4). DNase I is typically excluded by the bound protein for 5-6 nt at each end (29), so this 21 nt footprint probably corresponds to a 10 ± 1 nt region of protein:DNA contact. The structurally characterized co-crystal of R·PvuII involved 13mers with single nucleotide overhangs at each end (9). For comparison, the EcoRI endonuclease DNase I footprint was 18 nt long, with a 10 nt region contacted in the crystal structure (30,31). MboII endonuclease gave a 16 nt DNase I footprint (32) and HaeIII and HinP1I yielded footprints of 15 and 13 bp, respectively (30).

The length of the R·PvuII-protected region is also consistent with other observations. First, R·PvuII doesn’t cleave sites that are adjacent to Dam-methylated GATC sequences, even though the methylated adenine is outside the PvuII site (CAGCTGN6mATC) (33). Second, we found that a hairpin substrate showed strand-specific effects on cleavage if the loop was 8 bp from the CAGCTG site, but not if it was 15 bp away (data not shown). The hairpin oligonucleotides were designed with unequal arm lengths to reveal any strand preference in cleavage. When and only when the PvuII site was closer to the loop structure, the upper strand of the unmodified hairpin DNA was cleaved approximately three times faster than the lower strand (not shown).

When footprinting was performed on DNA that was labeled on the lower strand, the length and the characteristics of the footprint were the same as on the upper strand, including the hypersensitive site (not shown). However, the protected regions on the two strands appear to have 3[prime] overhangs relative to one another. This is indicative of minor groove binding (29), which is consistent with the crystal structure (9) and with predictions for endonucleases that generate blunt-end cleavages (34).


Figure 6. R·PvuII:DNA co-crystal structure is consistent with 5mC cleavage. A detail from the crystal structure of the R·PvuII:DNA complex includes one of the central G:C base pairs from the CAGCTG PvuII site and His83 and His84 from one of the two identical subunits of R·PvuII. The adjacent C:G base pair has identical interactions with His83 and His84 from the other subunit of R·PvuII. The structure in the center is from a complex containing unmethylated cytosine (9), which is the normal cleavage substrate; the structure was determined in the absence of Mg2+ so that no cleavage occurred. On the left, the structure shown is from a complex between R·PvuII and a PvuII site containing 5-iodoC (10); a methyl group has been modeled in place of the iodine atom and the purple circle indicates the van der Waals radius (which is 0.15 Å smaller than that of the iodine atom). In this structure, His84 was divided evenly between two conformations: one is close to that shown in the center drawing and the other has the histidine ring rotated downward (dark bonds). On the right, for comparison, is the normal protective methylation generated by M·PvuII. R·PvuII has very low affinity for N4-methylated PvuII sites (Rice and Blumenthal, in preparation) and this complex has not been crystallized. The structure shown is the G:5mC structure with the methyl group moved to the N4 position of cytosine and indicates the steric clash that would result.

The binding of R·PvuII to the plasmid fragment induced a very strong hypersensitivity to DNase I upstream of the PvuII site (Fig. 4, arrow). The EcoRI DNase I footprint also included a site of protein-induced DNase I hypersensitivity that was found to be outside the protected region (30). Hypersensitive sites have been correlated with DNA bending or other means of widening the major groove (29). However, the R·PvuII crystal structure showed no apparent bending of the 13mer DNA (9). While the behavior of DNA-binding proteins in solution and in complexes in crystals may differ in some cases, the hypersensitive site is at the end of the protected region on each strand and it is unclear how the bound protein would induce a bend at those points; protein-protein interactions between the bound R·PvuII and DNase I may be responsible for the observed hypersensitivity.

Binding of CAG5mCTG by R·PvuII

The effect of 5mC on the cleavage rate (Fig. 3) could be explained, in theory, by reduced binding, reduced kcat or reduced product release when 5mC is present. There was only a small negative effect of 5mC on binding as measured by the hypersensitive site generated when R·PvuII bound to the DNA in footprinting analyses, but in the competition assay the CAG5mCTG oligonucleotide competed with labeled substrate DNA no better than a non-specific (reversed site) oligonucleotide (Fig. 5). However, in comparison with the competition experiments, the footprinting reactions were performed in a different buffer, on a much longer DNA substrate and with an excess of enzyme over substrate. Furthermore, the DNA and R·PvuII were preincubated for 30 min prior to DNase I treatment, meaning that equilibrium binding was measured rather than kinetic behavior. Finally, the competition studies were carried out under reaction conditions in the presence of the required Mg2+ cofactor, while footprinting was carried out in the presence of Ca2+. If recognition and catalysis are coupled, as appears to be true for other restriction endonucleases (35), then a quantitative difference between competition and footprinting results would be expected. In the case of the EcoRV endonuclease, using a non-cleavable substrate, protein:DNA complexes formed in the presence of Ca2+ were found to be substantially more stable than complexes formed in the presence of Mg2+ (36).

The R·PvuII:DNA crystal structure is consistent with cleavage of CAG5mCTG sites

The R·PvuII:DNA crystal structure suggested functions for many of the amino acid sidechains in the protein. Most important for the study presented here is the hypothesis that His84, part of a histidine triplet located in the DNA recognition region, is responsible for sensing the methylation state of the relevant cytosine in substrate DNA (9). The C[beta] of His84 comes within 4.3 Å of the cytosine N4 in the co-crystal (in which the DNA was unmethylated; Fig. 6, center). N4 methylation could prevent C[beta] from coming so close and thus interfere with the hydrogen bonding that occurs between N[delta]1 of His84 and O6 of the paired guanine (Fig. 6, right). A study of random and site-directed mutants of R·PvuII included a mutant in which His84 was substituted by alanine (13). The H84A mutant gene was unable to transform a pvuIIM-deficient strain, indicating significant endonuclease activity in vivo, and the H84A protein had reduced but detectible cleavage activity on [lambda] DNA in vitro. N4 methylation of the central cytosine on the PvuII sequence still prevented cleavage by this mutant. The exact role of H84 thus remains unclear.

The R·PvuII:DNA co-crystal structure reveals considerable space around the 5-carbon of CAGCTG, though the closest amino acids, His83 and His84, are both constrained by hydrogen bonds to the DNA (9). The C5 position has considerably more space to accommodate a methyl group than N4 (Fig. 6, left), so 5mC would not have been expected to interfere with binding and cleavage by R·PvuII. While the structure of the R·PvuII complex with unmethylated DNA is consistent with our observation that CAG5mCTG sites are bound and cleaved, that structure does not explain why the cleavage is so much slower than with unmethylated DNA or why CAG5mCTG is such a poor competitor for unmethylated DNA under reaction conditions.

Our observation that R·PvuII was capable of cleaving CAG5mCTG had led us to suggest the use of CAG5-iodoCTG as a phasing derivative in R·PvuII:DNA co-crystals (37) since the van der Waals radius of iodine (2.15 Å) is close to that of CH3 (2.0 Å) (38). A crystal structure of R·PvuII complexed with DNA containing iodine at the C5 position of the methylatable cytosine (CAG5-iodoCTG) was recently solved to 1.9 Å resolution (10) and provides some insights into the effects of 5mC. First, the presence of 5-iodoC causes His84 to distribute into two distinct structural conformations. One conformation, shown by about half of the subunits, is the same as that seen in the complex with unmodified DNA, indicating that the presence of the iodine (and presumably the 5mC methyl group it mimics) is not sufficient to disrupt the approach of His84 to the central cytosine. In the other conformation, His84 is ~3.6 Å away from the iodine atom and is hydrogen bonded to the N4 atom of the methylatable cytosine (an interaction that could not occur with N4mC present). Whether this alternative conformation appeared in one subunit of each homodimer or instead appeared in both subunits of half the homodimers in the crystal could not be determined from the available data, but in either case the 5-iodoC clearly has a complex effect on the conformation of the His84 sidechain. Thus while much remains to be learned, the parallel structural and biochemical investigations are providing mutually consistent results regarding the effects of DNA methylation on R·PvuII conformation and activity.

ACKNOWLEDGEMENTS

We thank Drs Joan Dunbar and Ashok Bhagwat (Wayne State University) and Dr Xiaodong Cheng (Emory University) for critically reviewing the manuscript. We also thank Xiaodong Cheng and Dr John Horton (Emory University) for help with Figure 6 and for providing purified R·PvuII. This work was supported by the National Science Foundation under grant MCB-9631137.

REFERENCES

1. Gingeras,T.R., Greenough,L., Schildkraut,I. and Roberts,R.J. (1981) Nucleic Acids Res., 9, 4525-4536. MEDLINE Abstract

2. Blumenthal,R.M., Gregory,S.A. and Cooperider,J.S. (1985) J. Bacteriol., 164, 501-509. MEDLINE Abstract

3. Tao,T., Walter,J., Brennan,K.J., Cotterman,M.M. and Blumenthal,R.M. (1989) Nucleic Acids Res., 17, 4161-4175. MEDLINE Abstract

4. Athanasiadis,A., Gregoriu,M., Thanos,D., Kokkinidis,M. and Papamatheakis,J. (1990) Nucleic Acids Res., 18, 6434. MEDLINE Abstract

5. Tao,T. and Blumenthal,R.M. (1992) J. Bacteriol., 174, 3395-3398. MEDLINE Abstract

6. Tao,T., Bourne,J.C. and Blumenthal,R.M. (1991) J. Bacteriol., 173, 1367-1375. MEDLINE Abstract

7. Adams,G.M. and Blumenthal,R.M. (1995) Gene, 157, 193-199. MEDLINE Abstract

8. Athanasiadis,A., Vlassi,M., Kotsifaki,D., Tucker,P.A., Wilson,K.S. and Kokkinidis,M. (1994) Struct. Biol., 1, 469-475.

9. Cheng,X., Balendiran,K., Schildkraut,I. and Anderson,J.E. (1994) EMBO J., 13, 3927-3935. MEDLINE Abstract

10. Horton,J.R., Bonventre,J. and Cheng,X. (1998) Biol. Chem., 379, 451-458. MEDLINE Abstract

11. O'Gara,M., Adams,G.M., Gong,W., Kobayashi,R., Blumenthal,R.M. and Cheng,X. (1997) Eur. J. Biochem., 247, 1009-1018. MEDLINE Abstract

12. Gong,W., O'Gara,M., Blumenthal,R.M. and Cheng,X. (1997)Nucleic Acids Res., 25, 2702-2715. MEDLINE Abstract

13. Nastri,H., Evans,P.D., Walker,I. and Riggs,P. (1997) J. Biol. Chem., 272, 25761-25767. MEDLINE Abstract

14. Adams,G.M. and Blumenthal,R.M. (1997) Biochemistry, 36, 8284-8292. MEDLINE Abstract

15. Butkus,V., Petrauskiene,L., Maneliene,Z., Klimasauskas,S., Laucys,V. and Janulaitis,A.A. (1987) Nucleic Acids Res., 15, 7091-7102. MEDLINE Abstract

16. Kramarov,V.M. and Smolyaninov,V.V. (1981) Biokhimiya, 46, 1526-1529.

17. Brooks,J.E. and Roberts,R.J. (1982) Nucleic Acids Res., 10, 913-934. MEDLINE Abstract

18. Kang,S.C., Yoon,H. and Yoo,O.J. (1986) Korean Biochem. J., 19, 41-46.

19. McClelland,M., Nelson,M. and Raschke,E. (1994) Nucleic Acids Res., 22, 3640-3659. MEDLINE Abstract

20. Huang,L.H., Farnet,C.M., Ehrlich,K.C. and Ehrlich,M. (1982)Nucleic Acids Res., 10, 1579-1591. MEDLINE Abstract

21. Chung,C.T. and Miller,R.H. (1988) Nucleic Acids Res., 16, 3580. MEDLINE Abstract

22. Sambrook,J., Frisch,E.F. and Maniatis,T. (1989) Molecular Cloning: A Laboratory Manual, 2nd Edn. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

23. Pisa-Williamson,D. and Fuller,C.W. (1992) Comments, 19. US Biochemical, Cleveland, OH.

24. Howard,K.A., Card,C., Benner,J.S., Callahan,H.L., Maunus,R., Silber,K., Wilson,G. and Brooks,J.E. (1986) Nucleic Acids Res., 14, 7939-7951. MEDLINE Abstract

25. Brooks,J.E., Benner,J.S., Heiter,D.F., Silber,K.R., Sznyter,L.A., Jager-Quinton,T., Moran,L.S., Slatko,B.E., Wilson,G.G. and Nwankwo,D.O. (1989) Nucleic Acids Res., 17, 979-997. MEDLINE Abstract

26. Master,S.S. and Blumenthal,R.M. (1997) Mol. Gen. Genet., 257, 14-22. MEDLINE Abstract

27. Nelson,P.S., Papas,T.S. and Schweinfest,C.W. (1993) Nucleic Acids Res., 21, 681-686. MEDLINE Abstract

28. Kang,S. and Wells,R.D. (1994) J. Biol. Chem., 269, 9528-9532. MEDLINE Abstract

29. Suck,D. (1994) J. Mol. Recognit., 7, 65-70. MEDLINE Abstract

30. Fox,K.R. (1988) Biochem. Biophys. Res. Commun., 155, 779-785. MEDLINE Abstract

31. McClarin,J.A., Frederick,C.A., Wang,B.C., Greene,P., Boyer,H.W., Grable,J. and Rosenberg,J.M. (1986) Science, 234, 1526-1541. MEDLINE Abstract

32. Sektas,M., Kaczorowski,T. and Podhajska,A.J. (1995) Gene, 157, 181-185. MEDLINE Abstract

33. Chen,D., Liu,Q., Chen,X., Zhao,X. and Chen,Y. (1991) Nucleic Acids Res., 19, 5703-5705. MEDLINE Abstract

34. Anderson,J.E. (1993) Curr. Opin. Struct. Biol., 3, 24-30.

35. Pingoud,A. and Jeltsch,A. (1997) Eur. J. Biochem., 246, 1-22. MEDLINE Abstract

36. Engler,L.E., Welch,K.K. and Jen-Jacobson,L. (1997) J. Mol. Biol., 269, 82-101. MEDLINE Abstract

37. Balendiran,K., Bonventre,J., Knott,R., Jack,W., Benner,J., Schildkraut,I. and Anderson,J.E. (1994) Proteins, 19, 77-79. MEDLINE Abstract

38. Pauling,L. (1970) Chem. Br., 6, 468-472. MEDLINE Abstract

39. Zhang,B., Tao,T., Wilson,G.G. and Blumenthal,R.M. (1993) Nucleic Acids Res., 21, 905-911. MEDLINE Abstract

40. Tao,T., Walter,J., Brennan,K.J., Cotterman,M.M. and Blumenthal,R.M. (1989) Nucleic Acids Res., 17, 4161-4175. MEDLINE Abstract

*To whom correspondence should be addressed. Tel: +1 419 383 5422; Fax: +1 419 383 3002; Email: rblumenthal@mco.edu
+Present address: Department of Microbiology, State University of New York at Stony Brook, Stony Brook, NY 11790, USA

This page is run by Oxford University Press, Great Clarendon Street, Oxford OX2 6DP, as part of the OUP Journals
Comments and feedback: www-admin{at}oup.co.uk
Last modification: 29 Jan 1999
Copyright©Oxford University Press, 1999.
Add to CiteULike CiteULike   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us    What's this?


This article has been cited by other articles:


Home page
 Nucleic Acids ResHome page
I. Mruk, P. Rajesh, and R. M. Blumenthal
Regulatory circuit based on autogenous activation-repression: roles of C-boxes and spacer sequences in control of the PvuII restriction-modification system
Nucleic Acids Res., November 29, 2007; 35(20): 6935 - 6952.
[Abstract] [Full Text] [PDF]

Home page
 Nucleic Acids ResHome page
S. D. Pawlak, M. Radlinska, A. A. Chmiel, J. M. Bujnicki, and K. J. Skowronek
Inference of relationships in the 'twilight zone' of homology using a combination of bioinformatics and site-directed mutagenesis: a case study of restriction endonucleases Bsp6I and PvuII
Nucleic Acids Res., January 31, 2005; 33(2): 661 - 671.
[Abstract] [Full Text] [PDF]

Home page
 Nucleic Acids ResHome page
M. R. Rice and R. M. Blumenthal
Recognition of native DNA methylation by the PvuII restriction endonuclease
Nucleic Acids Res., August 15, 2000; 28(16): 3143 - 3150.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Print PDF (328K) Freely available
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (4)
Right arrowRequest Permissions
Right arrow Commercial Re-use Guidelines
for Open Access NAR Content
Google Scholar
Right arrow Articles by Rice, M. R.
Right arrow Articles by Blumenthal, R. M.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Rice, M. R.
Right arrow Articles by Blumenthal, R. M.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?