Cleavage of Mispaired Heteroduplex DNA Substrates by Numerous Restriction Enzymes 1 Cleavage of mispaired heteroduplex DNA substrates by numerous restriction enzymes

The utility of restriction endonucleases as a tool in molecular biology is in large part due to the high degree of specificity with which they cleave well-characterized DNA recognition sequences. The specificity of restriction endonucleases is not absolute, yet many commonly used assays of biological phenomena and contemporary molecular biology techniques rely on the premise that restriction enzymes will cleave only perfect cognate recognition sites. In vitro, mispaired heteroduplex DNAs are commonly formed, especially subsequent to polymerase chain reaction amplification. We investigated a panel of restriction endonucleases to determine their ability to cleave mispaired heteroduplex DNA substrates. Two straightforward, non-radioactive assays are used to evaluate mispaired heteroduplex DNA cleavage: a PCR amplification method and an oligonucleotide-based assay. These assays demonstrated that most restriction endonucleases are capable of site-specific double-strand cleavage with heteroduplex mispaired DNA substrates, however, certain mispaired substrates do effectively abrogate cleavage to undetectable levels. These data are consistent with mispaired substrate cleavage previously reported for Eco RI and, importantly, extend our knowledge of mispaired heteroduplex substrate cleavage to 13 additional enzymes. Introduction Restriction endonucleases are the DNA-cleaving components of restriction modification systems employed by bacteria, archaea and certain viruses of unicellular algae to protect against foreign, invading DNA (Pingoud and Jeltsch, 2001; Pingoud et al., 2005). Restriction endonucleases are classified broadly by subunit composition and cofactor requirement into types I, II, III and IV (Bickle and Kruger, 1993; Perona, 2002; Pingoud et al., 2005). The largest class of restriction endonucleases is type II, which is characterized by homodimeric or homotetrameric assemblies that generally require Mg as a cofactor and recognize and cleave close to or within a 4-8 base pair recognition sequence (Bickle and Kruger, 1993; Pingoud and Jeltsch, 2001; Perona, 2002; Pingoud et al., 2005). Most of the enzymes used in modern molecular biology laboratories -such as in assays of biological phenomena, cloning, mutation detection and site-directed mutagenesis protocols -are type II endonucleases (Pingoud et al., 1993; Roberts, 2005). The most valuable attribute of restriction endonucleases is the high degree of specificity with which they recognize and cleave DNA sequences (Roberts, 2005). Understanding the chemical and biophysical basis of this specificity has important biological and biomedical implications: The specificity of protein-DNA interactions underlies a myriad of regulatory and cell biological mechanisms. Crystal structures have been solved for several type II restriction endonucleases, and in some cases, mutant forms. Additionally, some are co-crystals with the enzyme bound to their cognate or noncognate substrates ((Roberts et al., 2007) and sources therein). Biophysical studies using crystallographic data, biochemical and kinetic studies, and molecular dynamics simulations have provided key insights into the proteinDNA interface -most notably with the EcoRI and EcoRV proteins -and have provided a foundational understanding of the biophysical basis of site specific interactions (Becker et al., 1988; Lesser et al., 1990; Lesser et al., 1993; Winkler et al., 1993; Engler et al., 1997; Jen-Jacobson, 1997; Horton and Perona, 1998b; Jen-Jacobson, 2000; Watrob et al., 2001; Horton et al., 2002; Parry et al., 2003; Kurpiewski et al., 2004; Hiller et al., 2005). Many type II restriction endonucleases are known to recognize a symmetrical site within DNA known as a palindrome, however, recently it is becoming increasingly appreciated that this class of enzymes is composed of proteins with a large degree of structural and functional diversity. The two-fold symmetrical relationship of the DNA-protein interface is consistent with the well-supported half-site rotational symmetry model of recognition (Kurpiewski et al., 2004). Functionally, type II restriction enzymes are typically a homodimer where the monomers interface symmetrically with each half-site of the recognition pattern (Alves et al., 1982). Although these enzymes have a high degree of specificity for their cognate recognition sequence, the specificity is not absolute. Extensive, pioneering biophysical studies have demonstrated the molecular basis of non-canonical DNA cleavage: the efficiency of non-canonical substrate cleavage can be predicted once the energetic penalties associated with the changes to the substrate are appreciated (Lesser et al., 1990). Restriction enzymes for which detailed studies of cognate and noncognate substrates have been performed, such as EcoRI, have revealed a dynamic range of dissociation and cleavage rate constants. Cleavage rate constants for noncognate substrates can vary from greater than that of the cognate substrate to a rate so slow that detectable cleavage is not generally observable (Thielking et al., 1990). For the enzymes EcoRV, EcoRI, BamHI the rate of noncognate substrate cleavage has been shown to be exquisitely dependent upon many variables, most notably: in vitro conditions, pH, salt concentrations (Engler et al., 1997), divalent ion concentrations (Thielking et al., 1992; Vermote and Halford, 1992; Vipond and Halford, 1995; Horizon Scientific Press. http://www.horizonpress.com Curr. Issues Mol. Biol. 11: 1-12. Online journal at http://www.cimb.org 2 Langhans and Palladino Engler et al., 2001; Sam et al., 2001; Kurpiewski et al., 2004), the presence of glycerol or other water occluding chemicals (Horton and Perona, 1998b; Sapienza et al., 2005; Sapienza et al., 2007), and the context of the site (adjacent nucleotides and the proximity to other recognition sites or DNA termini) (Alves et al., 1984; Engler et al., 1997; Jen-Jacobson, 1997; Horton and Perona, 1998a; Engler et al., 2001). These factors have widely variable thermodynamic contributions directly and indirectly to both the recognition and cleavage steps of restriction endonucleases (Lesser et al., 1990; Martin et al., 1999; Jen-Jacobson et al., 2000). While it is believed that most restriction enzymes share a common conserved functional protein core (Niv et al., 2007), there is very little primary amino acid sequence homology (Kovall and Matthews, 1999). Even among isochizomers, structural requirements for cleavage vary drastically (Wolfes et al., 1985). The lack of structural conservation suggests that mechanistic inferences between these enzymes should be made with extreme caution. Biophysical data pertaining to the mechanism of recognition and cleavage among type II restriction enzymes is currently limited to a few well-studied enzymes, making it exceptionally difficult to predict mechanistic characteristics or the rate of noncognate substrate cleavage of the many lesser-studied type II restriction enzymes. Heteroduplex DNA is not common in vivo: DNA mismatches are readily detected and repaired (Modrich, 1995). However, mismatch heteroduplex DNA is commonly found in vitro, especially when there is heterogeneity in the DNA and PCR is used to amplify the DNA. This had been elegantly demonstrated by T4 endonuclease VII sensitivity of mixed template PCR amplicons (Jenkins et al., 1999). T4 endonuclease VII has been shown to efficiently cleave mismatched DNA heteroduplexes (Golz et al., 1998a; Golz et al., 1998b). Assays utilized to quantify biological phenomena, such as the frequency of mutation, recombination or RNA editing, or common molecular biology techniques, such as sitedirected mutagenesis protocols, must address the issue of heteroduplex DNA when PCR amplification is employed. However, often assays utilize PCR and subsequently quantify the frequency of restriction endonuclease cleavage as a metric of a biological phenomenon, without controlling for heteroduplex DNAs. Numerous reports quantify restriction endonuclease cleavage in assays that ignore heteroduplex DNA formation or assume (sometimes this assumption is explicitly stated) that heteroduplex DNA will not be a viable substrate for double strand DNA cleavage (Iland and Todd, 1992; Paschen and Djuricic, 1994; Belcher and Howe, 1997; Judo et al., 1998; Serth et al., 1998; Jenkins et al., 1999). Protocols used to perform in vitro sitedirected mutagenesis, namely the unique site elimination (USE) method, also assume heteroduplex DNA will not be cleaved by restriction endonucleases (Deng and Nickoloff, 1992; Ling and Robinson, 1997; Clontech Laboratories, 2001). The assumption that a heteroduplex restriction sequence is not a viable substrate for cleavage is inconsistent with available structural and biophysical data suggesting that many non-cognate substrates may be efficiently cleaved in vitro, such as has been shown for the EcoRI endonuclease (Thielking et al., 1990). We investigated whether cleavage of mispaired heteroduplex DNA substrates was unique to EcoRI or was more widespread among commonly used restriction endonucleases. Both position and type of non-cognate change within the recognition sequence are used to empirically probe the DNA-protein interface for critical alterations that either permit or abrogate efficient cleavage. Results Heteroduplex DNA cleavage revealed by PCR assay Site selectivity of restriction endonucleases is a key property that has made these proteins extraordinarily useful in molecular biology. Despite their widespread use experimentally, a detailed understanding of the biophysical basis of this specificity has only been studied for a few enzymes, most notably EcoRI, EcoRV and BamHI (Perona, 2002). Due to the large degree of size and structural diversity of restriction endonucleases (Pingoud et al., 2005), inferences among enzymes should only be made with caution. In vivo, restriction enzyme recognition sites in DNA are almost always perfect cognate homoduplexes where the site (often a palindrome) exists on both strands of the DNA. Typically recognition of the two cognate DNA half-sites by the endonuclease (most often a dimer) results in specific double-stranded cleavage. In vitro, heteroduplex DNA can be quite common where one half-site may exist but the other half-site has a mismatch (thus the sequence is correct on only one strand). We asked whether such noncognate sites would be recognized by restriction endonucleases and double-strand cleavage would occur. We developed a PCR-based assay to test whether heteroduplex restriction enzyme recognition sites with a single nucleotide mispair could be cleaved by various endonucleases. The assay utilizes a forward primer encoding the restriction endonuclease recognition site or a non-cognate site bearing a single nucleotide substitution and a reverse primer to produce an ~ 300 based-pair amplicon from a plasmid template. The amplicons were quantified by absorption spectrometry, mixed equimolar, denatured, and annealed to ensure optimal formation of heteroduplex DNA (~50%). The annealed amplicons were utilized as substrate in a restriction endonuclease reaction and the products were resolved using agarose gel electrophoresis (Figure 1A). Control homoduplex samples with (+) and without (-) the cognate restriction recognition site were also tested. Because the heteroduplex substrate is ~ 1/4 homoduplex +, ~1/4 homoduplex -, and ~1/2 heteroduplex DNA, a cleavage pattern of 3/4 cutting and 1/4 resistant is consistent with efficient heteroduplex cleavage. The Eco0109I restriction endonuclease did not significantly cleave the heteroduplex DNA (Figure 1A). However, the other three enzymes examined, AhdI, PmeI and AlwNI, all produced a fragment pattern consistent with cleavage of some or all of the heteroduplex DNA (Figure 1A). Densitometry was performed on reactions performed in triplicate and Chi Square analyses were performed to determine whether the cleavage pattern (ratio of resistant to cleaved) matched that expected if only the homoduplex + DNA were cleaved. Only, Eco0109I produced such a pattern, whereas, AhdI, PmeI and AlwNI deviated significantly from that expected, suggesting the heteroduplex DNA also underwent significant doublestrand cleavage (Figure 1B). Cleavage of Mispaired Heteroduplex DNA Substrates by Numerous Restriction Enzymes 3 Figure 1: PCR-based assay for heteroduplex DNA cleavage assay. Oligonucleotide PCR amplicons ~300bp in length were amplified containing the recognition site (+) or with a single nucleotide substitution (-). These amplicons were used to generate heteroduplex DNA (HD) by annealing an equimolar mixture of + and – products. (A) Each enzyme was tested for cleavage of their recognition site (+), a mutated site (-), and heteroduplex substrates. All enzymes efficiently cleaved the positive control (+) and did not cut the mutated site (-). Mutations examined were AhdI #1I, Eco0109I #3I, PmeI #3V and AlwNI #1V mutations were examined. Nomenclature of mutation is the position number in the enzyme recognition sequence and type of change (transition = I and transversion = V) and is used throughout. (B) Quantification of the ratio of resistant to cleaved DNA is shown. Enzymes unable to cleave heteroduplex DNA are expected to produce a ratio of resistant to cleaved products equal to 3 (red dashed line), whereas, those efficiently cleaving heteroduplex DNA are expected to give a ratio of 0.33 (green dashed line). Only Eco0109 I produced a pattern that did not deviate significantly from that expected for not cutting heteroduplex (p > 0.95). AhdI, PmeI and AlwNI each deviated significantly from that expected for no heteroduplex cleavage (*** is p<0.001). Statistical analyses were Chi Square tests, n = 3 for each enzyme. Oligonucleotide assay of heteroduplex DNA cleavage We aimed to develop a more sensitive and quantitative, yet still non-radioactive, assay of mispaired non-cognate heteroduplex DNA cleavage using annealed oligonucleotides. The oligonucleotide assay utilizes custom synthesized and annealed oligonucleotides as the substrate. Reaction products are resolved using nondenaturing polyacrylamide gel electrophoresis (PAGE) and post stained with ethidium bromide for fluorescence analysis. We hypothesized the oligonucleotide heteroduplex assay would confirm our previous results with the PCR assay and yield more sensitive quantitative results, due to the homogeneity of the substrate. We examined three of the enzymes studied using the PCR assay (AhdI, Eco0109I, and PmeI) and ten additional commonly used restriction enzymes, including: BamHI, EcoRI, EcoRV, HindIII, PstI, SacI, SalI, SpeI, XhoI, XbaI, all for heteroduplex DNA cleavage with a single mispair. The oligonucleotide assay demonstrated SpeI, AhdI, PmeI and SacI restriction endonucleases cleave the heteroduplex substrate tested bearing a single noncognate mispair (Figure 2 and data not shown). XhoI, XbaI, EcoRI and SalI failed to efficiently cleave heteroduplex DNA (Figure 2A and 2B). Densitometry was performed to estimate cleavage efficiency with that of the homoduplex DNA bearing the site. In all cases, the minus enzyme lane did not reveal cleavage products and the homoduplex was robustly cleaved. The assays revealed SpeI, PmeI and SacI produced robust heteroduplex cleavage with PmeI and SacI approaching the efficiency of homoduplex cleavage in this assay (Figure 2C-E). Importantly, results obtained using the PCR-based heteroduplex DNA cleavage assay were confirmed with the oligonucleotide assay (Figure 2 E and data not shown). Position and mutation type effect on heteroduplex DNA cleavage Biophysical studies of certain restriction enzyme recognition sites have established a hierarchy of noncognate substrate viability from better than the cognate substrate to not a viable substrate at all (Perona, 2002). Mismatched substrates for the EcoRI enzyme are generally considerably better substrates than those 4 Langhans and Palladino bearing the alteration on both strands of the DNA. However, the specific alteration in mismatched sequences resulted in wildly varied cleavage rates spanning numerous orders of magnitude (Thielking et al., 1990). Additionally, the position of the alteration, type of alteration, destabilization of the duplex, perturbation of the DNA backbone and even flanking positions outside the recognition sequence have all been shown to affect substrate structure and or cleavage efficiency (Alves et al., 1984; Wenz et al., 1996; Engler et al., 1997; JenJacobson, 1997; Horton and Perona, 1998a). We examined the effect of position and mutation type on heteroduplex DNA cleavage while maintaining the context of the site constant. There are 18 possibly single mismatch substrates for a restriction endonuclease with a hexanucleotide recognition sequence when restricting the analysis to the common four DNA nucleotides. We reasoned that examining six mispaired substrates for each restriction endonuclease would efficiently estimate the frequency with which mispaired non-cognate substrates are cleaved. To initially probe the requirements of heteroduplex DNA cleavage activity, we generated alterations in each position, testing individually three transitions (I) and three transversions (V) in the recognition site (alternating where practical) and examined cleavage patterns using the oligonucleotide assay (Figure 3). Most of the enzymes examined produced readily detectable double-stranded cleavage fragments and, as predicted, none of the enzymes cleaved all of the mispaired heteroduplex DNA substrates tested (Figure 3 and Table 1). In all cases, double-strand cleavage of the homoduplex control was observed to be efficient (Figure 3). Importantly, none of the enzymes cleaved all of the mispaired substrates, ruling out contamination with a non-specif ic mismatch endonuclease. Interestingly, SpeI was found to robustly cleave heteroduplex DNA with transitions at positions 1, 3 or 5, but transversions at any position were cleaved less efficiently (Figure 3A). Although there was no strong position effect observed for SpeI, a mutation type preference was evident. In contrast, SacI revealed a strong position effect but no mutation type preference was observed (Figure 3B). SacI robustly cleaved heteroduplex DNA with transitions at positions 3 or 6 and transversions at positions 1 or 2, however, neither the transition at position 4 or the transversion at position 5 were viable substrates (Figure 3B). We extended our heteroduplex substrate position effect experimentation to include six other restriction endonucleases: XbaI, EcoRI, EcoRV, XhoI, BamHI, and HindIII. These experiments identified numerous additional sites where heteroduplex changes either permitted cleavage or abrogated cleavage (Figure 3C-H and Table 1). Some heteroduplex conditions, such as EcoRV 4I/+ and 6I/+, EcoRI 1V/+, XhoI 1V/+ and 2I/+, and BamHI 3I/+ and 6V/+, were efficiently cleaved, whereas, other heteroduplex substrates revealed minimal to no cleavage in these assays. Figure 2: Oligonucleotide heteroduplex DNA cleavage assay. Annealed oligonucleotides bearing the restriction recognition site (homoduplex = +/+) or lacking the site on one strand (heteroduplex = HD) were tested for DNA cleavage with various restriction enzymes. (A) SpeI #1I / + heteroduplex results in double stranded cleavage, producing a ~25 bp fragment. XhoI #6I/+, XbaI #2/+, and SalI #2I/+ heteroduplexes show no evidence of cleavage. (B-E) Quantified cleavage efficiency for homoduplex and heteroduplex DNAs from 3 or more independent reactions. (B) EcoRI #5V/+ heteroduplex is minimally cleaved. (C) Sac I #1V/+ heteroduplex is efficiently cleaved. (D) Spe I #1I/+ heteroduplex resulted in ~ 50% cleavage relative to homoduplex controls. (E) Pme I #6I/+ heteroduplex also cleaved efficiently. + and – indicate whether an enzyme was included in the restriction endonuclease reaction. +/+ substrate resulted in efficient cleavage for all enzymes examined. Arrows indicate cleavage fragment. Cleavage of Mispaired Heteroduplex DNA Substrates by Numerous Restriction Enzymes 5 Table 1: Summary of heteroduplex DNA cleavage activity for each restriction endonuclease. The positions examined for each type of non-cognate change are shown in parentheses and whether these were cleaved is indicated. G.C.T. (Glycerol concentration test) indicates whether heteroduplex cleavage requires or is stimulated by high glycerol concentrations. Enzyme Transition (I) Transversion (V) Assay(s) G.C.T. AlwNI N.E. Yes (1) PCR N.E. AhdI Yes (1) Yes (1) PCR & Oligonucleotide N.E. BamHI No (1,5) Yes (3) No (2,4,5) Yes (6) Oligonucleotide No Eco0109I No (3) No (3) PCR & Oligonucleotide N/A EcoRI Yes (5) No (2,4,6) Yes (1) No (3,5) Oligonucleotide Yes EcoRV Yes (3,4,6) No (2) Yes (5) No (1,3) Oligonucleotide No HindIII No (1,2,3,5) No (2,4,6) Oligonucleotide N/A PmeI Yes (5,6) Yes (5) PCR & Oligonucleotide N.E. PstI No (3) No (3) Oligonucleotide N/A SacI Yes (3,6) No (4) Yes (1,2) No (5) Oligonucleotide No SalI No (2) No (2) Oligonucleotide N/A SpeI Yes (1,3,5) No (2,4,6) Oligonucleotide No XbaI No (1,2,3,5) No (2,4) Yes (6) Oligonucleotide Yes XhoI Yes (2) No (4,6) Yes (1,6) No (3,5) Oligonucleotide No Note: The assay indicates the PCR-based heteroduplex cleavage (PCR) and / or the oligonucleotide heteroduplex cleavage (Oligo.) assays. ‡ = modest cleavage detected is 5-15%. * Heteroduplex cleavage at this position is glycerol dependent. No cleavage indicates less <5% of the appropriate fragment size was observed. N/A = not applicable. N.E. = not examined. Figure 3: Position effect of heteroduplex DNA cleavage. Position and mutation type affect heteroduplex cleavage differently for each enzyme examined. Three transitions (blue) and three tranversions (green) were examined for each enzyme. Heteroduplex substrates were examined at all six positions for: (A) SpeI, (B) SacI, (C) XhoI, (D) EcoRV, (E) EcoRI, (F) BamHI, (G) XbaI, and (H) HindIII. Each profile was tested two or more times for each enzyme and representative data are shown. Arrows indicate cleavage fragment. Red bullets highlight intermediate DNA fragment. 6 Langhans and Palladino Figure 4: Double mutant non-cognate cleavage analysis. (A) SpeI cleaves several single non-cognate1I/+, 3I/+, or 5I/+ heteroduplex DNAs (Figure 3). Heteroduplex substrates bearing all combinations of two of these non-cognate alterations in a single strand were examined as substrates for the Spe I endonuclease and failed to produce any detectable cleavage fragments. (B) Substrates bearing the homoduplex non-cognate sequence were examined for double-strand DNA cleavage with several enzymes using the oligonucleotide assay. +/+ and heteroduplex substrates produced double-stranded cleavage products (Arrow). Red bullets indicate intermediate sized fragment. (C) Substrates with two non-cognate changes, one alteration on each strand of DNA, were examined as substrates for the SpeI and SacI endonuclease. The single mutation heteroduplex substrates were efficiently cleaved, as before, however, substrates lacking the cognate signal on at least a single strand failed to produce significant DNA cleavage fragments. Several of the mispaired substrates produced an intermediate sized product in an enzyme dependent manner (Figure 3, red bullets). Control experiments with annealed oligonucleotide standards demonstrated these are consistent in size with single strand DNA cleavage. Kinetic studies with mismatched EcoRI substrates have shown a 400,000 fold range of cleavage rates between the canonical and certain non-canonical half-sites (Thielking et al., 1990), which is consistent with the apparent single strand cleavage observed for certain substrates using this oligonucleotide assay. Double mismatch substrate analysis in cis and trans orientations The finding that many single nucleotide mismatch substrates are susceptible to double strand DNA cleavage suggests the energetic penalties associated with the mismatch are minimal or at a minimum are often overcome under commonly used, multiple turnover, in vitro conditions. Heteroduplex DNAs with individual transition mismatches at positions 1, 3, or 5 in the SpeI recognition sequence resulted substrates susceptible to cleavage using the oligonucleotide assay (Figure 3). Since each mismatch mutation resulted in cleavage, we asked whether any combination of the two transitions in cis would produce detectable cleavage products. As before, the 3I/+ control reaction produced robust cleavage, however, none of the double missense substrates in cis produced significant cleavage fragments (Figure 4A). For several enzymes that produced robust cleavage of specific mismatch heteroduplex substrates, we examined substrates bearing this mismatch alteration in trans to determine their viability as substrates. In all cases the positive control reactions produced the predicted cleavage fragments, however, the double mismatch alteration in trans abrogated cleavage (Figure 4B). We examined the panel of six mutations all in trans with the SpeI 3I alteration and in trans with the SacI 1V alteration to determine whether any of these combinations resulted in a viable substrate using the oligonucleotide assay. Although the single mismatch positive controls produced robust cleavage products, none of the double mutations in trans resulted in significant cleavage (Figure 4C). Figure 5: Effect of glycerol on heteroduplex substrate cleavage. Glycerol concentrations 2-14% were examined for the Sac I 2V/+, Spe I 3I/+, Xho I 2I/+, Bam HI 3I/+, PmeI 6I/+, EcoRV 6I/+, EcoRI 1V/+, and Xba6V/+ heteroduplex DNA substrates. SacI, SpeI, XhoI, BamHI, PmeI, and EcoRV efficiently produced double-stranded DNA fragments independent of glycerol concentration. Arrows indicate cleavage fragment. In contrast, EcoRI cleavage of the 1V/+ heteroduplex DNA was strictly glycerol dependent and XbaI cleavage of the 6V/+ substrate did not require high glycerol but was stimulated by glycerol. The total unit hours used per picomole of substrate is provided. Cleavage of Mispaired Heteroduplex DNA Substrates by Numerous Restriction Enzymes 7 Glycerol effect on mismatch heteroduplex cleavage. We examined the effect of glycerol on cleavage of several single nucleotide mismatch heteroduplex DNAs. Glycerol and other water occluding chemicals can enhance cleavage of non-cognate substrate cleavage by some restriction endonucleases. Reduced specificity of some restriction enzymes under conditions of high glycerol may result from the dramatic reduction in the effective reaction volume and or by the removal of water molecules from the enzyme-substrate interface that are required for specificity. We performed our oligonucleotide assay with varying concentrations of glycerol (2-14%) to determine whether heteroduplex DNA cleavage required or was stimulated by high glycerol concentrations. The reaction volume was increased 3 fold for these experiments and samples were taken at multiple time points, reducing the total reaction time 25-85%. In all cases, the most informative partial digestion is shown (Figure 5). Only the EcoRI 5I/+ and 1V/+ sites were found to require high glycerol concentration to produce detectable cleavage. The XbaI 6V/+ DNA did not require high glycerol concentration but glycerol appeared to stimulate heteroduplex DNA cleavage of this substrate. The remaining enzymes tested, including SacI, SpeI, XhoI, BamHI, PmeI and EcoRV, neither required high glycerol for heteroduplex DNA cleavage nor were detectably stimulated by glycerol for the substrate examined (Figure