Self-catalyzed site-specific depurination of guanine residues within gene sequences

Abstract

A self-catalyzed, site-specific guanine-depurination activity has been found to occur in short gene sequences with a potential to form a stem-loop structure. The critical features of that catalytic intermediate are a 5′-G-T-G-G-3′ loop and an adjacent 5′-T·A-3′ base pair of a short duplex stem stable enough to fix the loop structure required for depurination of its 5′-G residue. That residue is uniquely depurinated with a rate some 5 orders of magnitude faster than that of random “spontaneous” depurination. In contrast, all other purine residues in the sequence depurinate at the spontaneous background rate. The reaction requires no divalent cations or other cofactors and occurs under essentially physiological conditions. Such stem-loops can form in duplex DNA under superhelical stress, and their critical sequence features have been found at numerous sites in the human genome. Self-catalyzed stem-loop-mediated depurination leading to flexible apurinic sites may therefore serve some important biological role, e.g., in nucleosome positioning, genetic recombination, or chromosome superfolding.

Depurination in DNA has been recognized as a spontaneous form of intracellular DNA damage that affects both G and A residues in an essentially random way (1). In mammalian cells, such damage has been estimated to occur with a k_obs of 3 × 10⁻⁹ min⁻¹ (2). A complex and elaborate cellular DNA repair machinery has evolved to repair apurinic sites (3, 4). It has long been known that some G residues are more prone to depurination than others (5, 6), and such depurination hot spots have been associated with elevated mutation rates (6, 7). Notwithstanding these insights, notions to account for DNA depurination that go beyond recognition that hydrolysis of the purine–deoxyribose glycosyl bond is acid-catalyzed have not been forthcoming.

In the course of studies using a 29-residue single-stranded coding strand fragment encompassing the mutation site of the sickle cell β-globin gene (8–10), we found that this 29-nt oligomer as well as its shortened 18-nt segment self-catalyze site-specific depurination of a singular G residue adjacent to the mutation site. In this report, we identify the precise location of the depurination site, the reaction mechanism and its immediate products, the structure of the catalytic intermediate, and some of the sequence requirements and their tolerance for variation within the intermediate. We also note the wide distribution of such putative self-depurinating sequences in the human genome.

Results

Specific Backbone Cleavage Is the Result of Site-Specific Depurination.

It is as a consequence of slow spontaneous backbone cleavage at apurinic sites that the self-catalyzed site-specific depurination we report here was discovered. At the outset, it was observed that this cleavage occurs whether or not 10⁻³ M EDTA is present, indicating that divalent cations are not required for the cleavage. Incubation of the highly purified and homogeneous 29-nt oligomer D-1 at 22°C (to allow depurination to proceed) spontaneously yielded a 19-nt fragment and an ≈9- to 10-nt group of four fragments revealed by labeling their 5′ ends with ³²PO₄ (Fig. 1). However, when 3′-labeling with terminal deoxytransferase (11) was performed instead, only the ≈19-nt band was visible, whereas the shorter group of bands was barely visible (data not shown). The presence of the several bands refractory to terminal deoxytransferase, with mobilities corresponding to ≈9–10 nt, is consistent with depurination that results in an equilibrium between C1′ hemiacetal and C1′ aldehyde sugars at the apurinic site, accounting for two of the bands. Subsequent loss of either the terminal phosphate only or both the phosphate and sugar, each resulting in successively increasing fragment mobility, would account for the other two. Piperidine treatment at 75°C accelerates such backbone cleavage by base-catalyzed β-elimination (12, 13), thereby allowing detection and quantitation of the accumulated apurinic sites.

Fig. 1.

Localization of the cleavage site. A ³²P-labeled 8-nt oligomer complementary to the 3′ end of D-1 was used as a primer in a Sanger sequencing reaction. The appearance of several ≈9- to 10-nt cleavage products in the D-1 template lane labeled at the 5′ end with ³²P after piperidine treatment is consistent with site-specific depurination followed by β-elimination.

The initial backbone cleavage site was located by comparing the large cleavage product with a sequencing ladder on denaturing PAGE (Fig. 1). The large cleavage product band corresponded in mobility to a DNA polymerase product resulting from incorporation of 2′,3′-dideoxyadenosine opposite the sickle cell mutant T residue immediately downstream from the depurinated G. This band appeared equally intense in all four residue lanes because of significant presence of the 19-nt cleavage product in the D-1 template, as revealed by polynucleotide kinase labeling. The next band was also equally intense in the four lanes, consistent with the proclivity of DNA polymerase to insert any residue opposite an apurinic site (14).

The location of the depurination site and the presence of a 5′ phosphate on the 19-nt product of spontaneous cleavage were confirmed by successful ligation of a synthetic 10-nt oligomer bearing a 3′-OH and a 5′-phosphate, and the gel-purified 19-nt spontaneous cleavage product after their annealing to the complementary 29-nt anti-D-1 strand (see Table 1 and Fig. 2). Ligation can succeed only when the two fragments annealed with the 29-nt complementary oligomer are immediately adjacent. Because the synthetic 10-nt oligomer had a 3′-OH, the 19-nt fragment must bear a 5′-phosphate for ligation. In fact, ligation produced a full-size D-1 strand, as well as a D-1 dimer, a product of blunt-end ligation of two duplexes (Fig. 2, lane 1). Alkaline phosphatase treatment of the 19-nt fragment before annealing resulted in only a trace of full-size D-1 (Fig. 2, lane 3). Because the integrity of the 19-nt fragment was unaffected by the phosphatase treatment (data not shown), it must be the removal of its 5′-phosphate that prevents ligation.

Table 1.

Some deoxyoligonucleotides used in the study

Deoxyoligonucleotide	Sequence
Anti-D-1	3′-GACTGAGGACACCTCTTCAGACGGCAATG-5′
D-1	5′-CTGACTCCTGTGGAGAAGTCTGCCGTTAC-3′
DH-1	5′-CTGACTCCTGAGGAGAAGTCTGCCGTTAC-3′
D1-T	5′-CTGACTCCTTTGGAGAAGTCTGCCGTTAC-3′
D1-A	5′-CTGACTCCTATGGAGAAGTCTGCCGTTAC-3′
D1-C	5′-CTGACTCCTCTGGAGAAGTCTGCCGTTAC-3′

Fig. 2.

Ligation of the 5′-³²P-labeled 19-nt spontaneous cleavage product and the 10-nt synthetic fragment on the complementary anti-D-1 template. Lane 1, ligation without prior phosphatase treatment. This denaturing gel indicates that the unlabeled 19-nt fragment has a 5′-phosphate and so is successfully ligated with the 10-nt oligomer on the complementary template. Lane 2, ³²P-labeled 10-nt oligomer. Lane 3, ligation after alkaline phosphatase treatment of the purified 19-nt fragment yields almost no ligated D-1. Lane 4, ³²P-labeled fresh D-1 stock.

A Guanine Residue Is Required 5′ to the Cleavage Site.

To determine the sequence requirements immediately adjacent to the cleavage site that are essential for the depurination, several 29-nt sequence variants (Table 1) were compared. The site-specific depurination occured only when the residue 5′ to the cleavage site was G (Fig. 3). Its replacement by a pyrimidine eliminated cleavage, and only a trace of the cleavage product was visible for the A replacement. Cleavage was significantly less for the wild-type oligomer DH-1, indicating that the nature of the immediate downstream residue is also important. It is worth noting that all of these 29-nt oligomers had at least six other G residues in their sequences (seven in case of D-1 and DH-1, the sickle cell and wild-type β-globin gene fragments, respectively); yet none of them displayed any detectable depurination. Taken together, the results in Figs. 1 and 3 show that the depurination is very site-specific.

Fig. 3.

Piperidine cleavage of 29-nt sequence variants of the D-1 fragment of the sickle cell β-globin gene coding strand after 48-h incubation at 22°C (pH 5.0). No cleavage was evident when a pyrimidine residue replaced the specifically depurinated G residue. Only a trace of cleavage product was evident with an A substitution.

Free Guanine Is One Initial Product of the Catalysis.

With depurination presumed to be the primary event leading to backbone cleavage, direct evidence was sought for free guanine and the resultant apurinic site. Fig. 4 compares HPLC profiles for incubated D-1, D1-T [which does not undergo piperidine-induced cleavage and therefore does not lose its base (Fig. 3)], and a mixture of D1-T and deoxyguanosine (dG) as a marker (neither dG nor guanine binds to the C18 column), with the molar ratio of dG:D1-T similar to that expected from the catalytic depurination of oligomer D-1. The small, immediately eluted peak in the D1-T/dG mixture was absent in the elution profile of the D1-T sample alone, but such a peak was present in the elution profile of incubated D-1. When HPLC was conducted under highly acidic conditions (10% trichloroacetic acid), this small peak displayed, in addition to UV absorption, strong blue fluorescence characteristic of free guanine. Thus, one primary product of self-catalytic depurination was defined.

Fig. 4.

HPLC profiles of incubated D1-T, a mixture of D1-T/dG, and incubated D-1 from a C18 reverse-phase column eluted with a 10–60% acetonitrile gradient (diagonal dotted line), started at 3 min, and ended at 13 min of elution in 25 mM sodium phosphate (pH 7.0). Strand concentrations were 0.05 mM, and free dG concentration was 0.01 mM. Neither dG nor guanine binds to the column and is eluted within the first 4 min. The slight difference between the elution profiles of incubated D-1 and of the D1-T/dG mixture is due to the deoxyribose moiety of the dG (deoxynucleoside) marker present in the “artificial” mixture and absent in the D-1 sample, because it is free guanine that is accumulated during the incubation.

Accumulation of Apurinic Sites Revealed by Piperidine Treatment Is Paralleled by Apurinic Endonuclease Cleavage.

Fig. 5illustrates accumulation of apurinic sites monitored both by piperidine cleavage (Fig. 5A) and by cleavage with an apurinic endonuclease (Fig. 5B). The apurinic endonuclease activity of exonuclease III (Exo III) was used for this purpose. When an 18-nt subfragment of the D-1 strand, D(-2), which is depurinated at the same rate as D-1 (Table 2), was incubated at pH 5 and 22°C for various times and then digested with Exo III, a 10-nt band, the equivalent of the 19-nt 3′-end digestion product of D-1, appeared on the gel. As in the case of piperidine cleavage, the digestion site was unique; but, unlike piperidine cleavage, which produces several short fragments, the Exo III 7-nt digestion product at the 5′ side of the 18-nt D1(-2) was also unique (data not shown). The slope of the initial kinetics of cleavage was the same whether by piperidine or Exo III (Fig. 5C), indicating that they both reflect the rate of accumulation of apurinic sites. Hence, the self-catalyzed depurination gave rise to the apurinic sites that are cleavable by either piperidine treatment or apurinic endonuclease.

Fig. 5.

Kinetics of apurinic site accumulation monitored by piperidine or apurinic endonuclease cleavage. (A) Denaturing PAGE analysis of cleavage products of 3′-labeled D-1 samples incubated in 10 mM sodium cacodylate (pH 5) at 22°C, with or without subsequent piperidine treatment at 75°C for 30 min. Note that the cleavage products are highly homogeneous. (B) Exo III digestion at the apurinic site produces a 10-nt cleavage product of the 18-nt D-1 subfragment D(-2) labeled at the 3′ end with terminal deoxytransferase. The band immediately below D(-2) is the n-1 oligomer that was not fully removed during PAGE purification. (C) Kinetics of self-catalyzed, site-specific depurination revealed by Exo III digestion at apurinic sites (squares), piperidine cleavage (triangles), and nonspecific cleavage by piperidine (diamonds). k_obs for the depurination rate revealed by Exo III digestion coincides with that determined from piperidine-induced cleavage. k_obs for the background-nonspecific digestion rate is <10⁻⁶ min⁻¹. (D) Kinetics of self-catalyzed site-specific depurination of D-1 at three different concentrations over two orders of magnitude. The apurinic sites were cleaved with piperidine at 70°C for 30 min, and the yield of 19-nt cleavage product was determined after PAGE. Diamonds, 5 × 10⁻⁷ M D-1; squares, 5 × 10⁻⁶ M D-1; triangles, 5 × 10⁻⁵ M D-1; filled circles, anti-D-1.

Table 2.

Reduction in the rate of self-catalytic depurination when D-1 sequence is shortened

Oligomer	Sequence	kobs, min−1
D-1	5′-CTGACTCCTGTGGAGAAGTCTGCCGTTAC-3′	2 × 10−4
D-20	5′-CTGACTCCTGTGGAGAAGTC-3′	2 × 10−4
D(-2)	5′-GACTCCTGTGGAGAAGTC-3′	2 × 10−4
D-19	5′-CTGACTCCTGTGGAGAAGT-3′	4 × 10−5
D(-3)	5′-ACTCCTGTGGAGAAGTC-3′	3.6 × 10−5
D-18	5′-CTGACTCCTGTGGAGAAG-3′	1 × 10−5
D-17	5′-CTGACTCCTGTGGAGAA-3′	<10−6
D-14	5′-CTGACTCCTGTGGA-3′	<10−6
D(-6)	5′-CCTGTGGAGAAGTC-3′	<10−6

The boldface G residue is depurinated, and residues that form stem base pairs are underlined.

It is also noteworthy that the kinetics of depurination, as measured by the yield of piperidine cleavage products, was independent of strand concentration over 2 orders of magnitude (Fig. 5D), with k_obs = 2 × 10⁻⁴ min⁻¹. This result is consistent with first-order kinetics and indicates that the catalytic intermediate is not a multistranded structure; rather, an intramolecular structure is critical for the self-catalytic depurination.

Self-Catalytic Depurination Correlates with Stability of a Stem-Loop Formed by D-1.

The rate of catalytic depurination correlates with the ability of the oligomer to form a stem-loop with a 7-bp stem [including an A·C⁺ mismatch (15)] closed off at one end by a 4-nt loop, 5′-G-T-G-G-3′. It is the 5′-G in this loop that undergoes the site-specific self-catalytic depurination.

Table 2 indicates how the rate of this depurination changes with shortening of the oligonucleotide. Removal of 2 nt from the 5′ end and 9 nt from the 3′ end of the 29-nt D-1 gene fragment [D(-2)] had no effect on the depurination rate, apparently because it had no consequence for stem formation. However, further shortening reduced the number of possible base pairs in the stem of the hairpin, obviously leading to its destabilization and, with that, reduction in the depurination rate as well. Thus, when D-1 was shortened on either end, so that only 6 bp of the stem could form [D(-3) and D-19], the depurination rates fell ≈5-fold; with only 5 bp possible, there was an ≈20-fold reduction (D-18). No depurination above the background rate was detected for D1(-6) or D-17, which have a potential to form only 3 or 4 bp in the stem (one of them mismatched). Thus, the reduction in depurination rate correlated with the destabilization of the hairpin structure.

Formamide destabilizes base pair hydrogen bonds, reducing the T_m value of duplex DNA by ≈0.65°C for every 1% in the solvent (16). Because the T_m value for the D-1 stem-loop is 42°C and a formamide concentration of 50% reduces T_m by ≈30°C, it is not surprising that the self-catalytic depurination activity is completely lost at 22°C.

The effect of temperature on the depurination rate is also informative. The rate of spontaneous acid-catalyzed depurination should follow the usual temperature dependence, doubling with every 10°C rise; the background depurination rate at 22°C of 3 × 10⁻⁹ min ⁻¹ was determined by extrapolating rate data measured at 70°C based on this assumption. In contrast, the site-specific self-catalytic depurination rate showed an inverse dependence on temperature (Fig. 6). Thus, after 45 h of incubation at 22°C, 70% of D-1 was cleaved at the specific depurination site by piperidine treatment, but only 40% was cleaved after incubation at 43°C (just above T_m, where half the molecules would be in the hairpin form), and only 16% was cleaved at 65°C. This inverse dependence must reflect the thermal denaturation of the stem-loop structure required for the self-catalysis.

Fig. 6.

Temperature-dependence of site-specific and nonspecific spontaneous G residue depurination. Whereas spontaneous depurination progressively increases between 22°C and 65°C after 45 h of incubation, the temperature-dependence of the site-specific depurination is inverted. The reduction of specific G-residue depurination at higher temperatures reflects denaturation of the catalytic intermediate stem-loop structure, such that the level of its depurination is the same as for the other G residues.

The spontaneous depurination leading to nonspecific cleavage at other sites within the same molecule is also illustrated in Fig. 6. That process accelerated with increasing temperature, so that depurination was the same (14–16%) for all of the G residues at 65°C. Taken together, these results are consistent with a stem-loop being the catalytic intermediate for the self-catalytic depurination reaction.

The Stem-Loop Structure Is Confirmed by the Mung Bean Endonuclease Digestion Pattern.

A diagram of the putative stem-loop is shown along with the 18-nt subfragment cleavage pattern obtained with single-strand-specific mung bean endonuclease in Fig. 7. Three major cutting sites are indicated in Fig. 7A, and their fragmentation pattern is shown in Fig. 7B, lanes 2 and 3. These cleavage sites surround the apurinic site localized by apurinic endonuclease (Fig. 7B, lane 4). The reason for much weaker digestion at the 3′ end of the loop, where there are two other G residues, may be their strong tendency to stack on each other and on the following 3′-base-paired A residue. Regardless, the mung bean endonuclease digestion pattern clearly differentiates between the putative loop and stem residues, thereby confirming the stem-loop structure.

Fig. 7.

The stem-loop formed by the D-2 subfragment, and its digestion with mung bean endonuclease. (A) The minimal fully active self-depurinating oligomer folds into a stem-loop structure with one A·C⁺ mismatch in the stem. The dotted arrow indicates the site of G-depurination, which is then cleaved by Exo III. The solid arrows indicate phosphodiester bonds digested by the single-strand-specific mung bean endonuclease. (B) PAGE analysis of mung bean endonuclease and Exo III digests (pH 5, 22°C) of the minimal fully active oligomer D1(-2) after labeling at the 3′ end and a 24-h incubation at 22°C to allow depurination to proceed. Lane 1, control (no added enzyme); lanes 2 and 3, mung bean nuclease digests after 1 and 5 min of incubation, respectively; lane 4, Exo III digestion at the apurinic site.

Effects of Sequence Variation Within the Stem-Loop Define Requirements for the Self-Catalytic Depurination.

Several 18-nt sequence variants of the stem closed by the 5′-G-T-G-G-3′ loop were investigated: an “inverted” variant [D(-2)inv] in which the stem strands were exchanged; a “perfect” variant [D(-2)p] in which the A·C⁺ mismatch was replaced by a G·C base pair; a “changed” variant [D(-2)ch] with an arbitrary stem duplex sequence; and a “disrupted” variant [D(-2)d] in which the number of potential mismatches in the stem should prevent duplex formation at room temperature. As shown in Table 3, k_obs for the perfect stem sequence was approximately twice that of the natural sickle cell β-globin-derived sequence D(-2), suggesting that a more stable stem enhances the catalytic activity. The inverted variant showed a substantial loss of activity even though a complementary base-paired duplex structure was maintained. However, the changed oligomer, in which only the first base pair adjacent to the loop (5′-T·A-3′) and the A·C⁺ mismatch were conserved, retained full activity. Thus, this T·A base pair appears to be essential for the self-catalytic activity. Moreover, when the stem was fully complementary, just 5 bp were sufficient for full catalytic activity (D14p). Not surprisingly, the disrupted variant exhibited no activity. These results define the required consensus sequence and structural elements.

Table 3.

Effect of the stem-region sequence on the self-catalytic depurination rate

Strand	Sequence	kobs, min−1
D(-2)p	5′-GACTCCTGTGGAGGAGTC-3′	4 × 10−4
D14p	5′-CTCCTGTGGAGGAG-3′	4 × 10−4
D(-2)	5′-GACTCCTGTGGAGAAGTC-3′	2 × 10−4
D(-2)ch	5′-GGACCCTGTGGAGAGTCG-3′	2 × 10−4
D(-2)inv	5′-CTGAAGAGTGGTCCTCAG-3′	3 × 10−5
D(-2)d	5′-GACTCCTGTGGGAAGGTC-3′	<10−6

The boldface G residue is depurinated, and residues that form stem base pairs are underlined.

Self-Catalytic Depurination Under Physiological Conditions.

Although the results presented above were obtained at pH 5 and low salt concentration, the catalytic activity, albeit slightly reduced, was retained under essentially physiological conditions. Fig. 8 shows site-specific depurination kinetics revealed by piperidine cleavage of D(-2)ch after incubation at 37°C in 0.1 M NaCl/5 mM MgCl₂/20 mM Tris-HCl, pH 6.9/20% PEG 1000 to mimic intranuclear macromolecular crowding (17). Both the gel and its quantitation clearly demonstrate the occurrence of the site-specific self-catalytic depurination; its k_obs is 7 × 10⁻⁶ min⁻¹, ≈30 times slower than at pH 5, but still ≈3 × 10³ faster than spontaneous background depurination. It is important to note that for the D(-2)p strand, in which the A·C⁺ mismatch in the stem is replaced with a G·C base pair, the reduction in the depurination rate under physiological conditions is only ≈15 times (data not shown), indicating that the acidic milieu required for maximum activity of the original sequence is due partially to the stabilization of the A·C⁺ mismatch.

Fig. 8.

Self-catalytic site-specific G-depurination under physiological conditions. (A) Denaturing PAGE analysis of piperidine cleavage products of 3′-³²P-labeled D(-2)ch incubated in 20 mM Tris-HCl/0.1 M NaCl/0.05 M MgCl₂ (pH 6.9) at 37°C. (B) Kinetics of self-catalyzed, site-specific depurination revealed by piperidine cleavage at the specific site (diamonds) and nonspecific depurination sites (circles).

Potential Depurination Sites Are Abundant in the Human Genome.

Preliminary data mining was undertaken by using a Perl search engine to find sequences in the human genome with a potential to undergo self-catalytic G-depurination. We have already identified >50,000 sites, distributed among all human chromosomes, as having a central 5′-G-T-G-G-3′ (loop) sequence, an adjacent 5′-T·A-3′, and a capacity to form a stem of at least four additional complementary base pairs. Two of these sequences, both 14 nt in length, have been experimentally tested so far, and each was found to possess self-depurinating activity with a k_obs of 5 × 10⁻⁴ min⁻¹ at pH 5.

Discussion

We have demonstrated a site-specific self-G-depurinating activity in a short DNA fragment of the sickle cell β-globin gene that is ≈3 × 10³ faster under physiological conditions (≈10⁵ faster at pH 5) than the spontaneous depurination rate (2). In this connection, the pH optimum shifts toward neutrality when the A·C⁺ base pair of the stem is replaced by a G·C base pair.

Spontaneous depurination of DNA, a major type of endogenous DNA damage, is thought to be mediated by protonation of a G residue at N-7, with an intrinsic pK_a of ≈2.3. Presumably, such protonation is also implicated in the self-catalyzed depurination, probably facilitated in part by intracellular macromolecular crowding known to alter acid-base equilibria (17). Because only one of the three G residues in the “catalytic loop” is depurinated, it is likely that some subtle aspect of the stem-loop structure contributes as well.

We have shown that the self-catalytic depurination rate correlates with the ability of a single-stranded DNA fragment to form a stable hairpin with a 5′-G-T-G-G-3′ loop and the adjacent 5′-T·A-3′ base pair of the stem. The remaining sequence requirement is only for an additional number of complementary base pairs to stabilize the loop residues stacked on the T·A base pair. With increasing stability of the stem, helix breathing must be reduced, increasing the time average over which the catalytic conformation is maintained, hence a faster rate of depurination.

The critical sequence features for this G-depurinating activity are present in numerous natural gene sequences. The activity is self-catalytic and requires no cofactors, and the consensus stem-loop structure is far smaller than the only reported site-specific G-depurinating deoxyribozyme obtained by in vitro selection (18). The latter activity is contained within a complex three-stem-loop structure, requires divalent cations, and acts on a singular G residue in the three-residue loop of a stem-loop that is bound to the deoxyribozyme by complementary base pairing. Thus, this catalyst depurinates a G residue of a deoxyoligomer substrate, acting like a true enzyme with turnover, unlike our activity, which does not represent catalysis in the enzymatic sense because turnover is not possible.

Because the self-depurinating activity is associated with a single-stranded deoxyoligomer, it might be viewed as biologically irrelevant. However, superhelical stress in vivo induces formation of stem-loops in duplex DNA at inverted repeats (19, 20), and stems as short as 5–7 bp easily extrude under physiological superhelical stress (21).

Many sites with a potential to form stem-loops are known to be hypermutagenic (22, 23). The ability of some stem-loop structures to depurinate a specific loop residue could explain such mutagenicity. But if the result of depurination were merely to mutilate or otherwise damage DNA, then we would expect that evolution would have found a way to reduce the number of such sites (by making use of alternative codons that would not disturb encoded proteins but would disrupt stem base pairing). Because this is clearly not the case, formation of apurinic sites may serve some functional role. Thus, apurinic sites also slightly destabilize DNA duplexes (24) and increase local DNA flexibility (25), which may, for example, facilitate formation of open DNA–protein complexes and/or play a role in DNA packaging in viruses, in nucleosome formation or positioning, or in facilitating genetic recombination.

With the discovery of a site-specific, self-catalytic, G-depurinating activity intrinsic to numerous DNA sequences, we are alerted to the possibility of yet other self-catalytic activities contained within genomic DNA.

Materials and Methods

Deoxyoligonucleotides.

Deoxyoligonucleotides were purchased from Integrated DNA Technologies (Coralville, IA), purified to homogeneity by denaturing PAGE, and recovered from gel slices by soaking overnight in 0.1 M sodium phosphate (pH 7.0). Final purification was accomplished by 100% acetonitrile elution from a C18 Sep-Pak reverse phase column (Waters) followed by spin evaporation. Oligomer concentration was then adjusted spectrophotometrically in Aldrich ACS reagent-grade water.

3′-End Labeling.

3′-End labeling was performed by using terminal deoxytransferase (USB Corp.) (5 units) on ≈5 pmol of purified oligomer with ≈5 pmol of [α-³²P]dideoxyATP (Amersham Pharmacia Biotech) (13). The labeled oligomers were purified by denaturing PAGE before use.

Kinetics of Depurination.

Oligomers were incubated at 22°C in 10 mM sodium cacodylate (pH 5.0) (unless otherwise noted), and aliquots were frozen on dry ice at appropriate intervals. To reveal apurinic sites, aliquots were treated with 0.1 M piperidine for 30 min at 75°C to induce base-catalyzed backbone cleavage (12, 13). Either a trace amount of end-labeled oligomer was mixed with an unlabeled stock of appropriate concentration before incubation or the aliquots were labeled after the piperidine treatment (both methods yield the same depurination rates). Radioactivity of intact oligomer, its 3′-end cleavage product, and nonspecific cleavage at other G residues were quantitated by using a Molecular Dynamics PhosphorImager, and these data were used for the kinetic analysis. Each experiment was performed at least three times, and the kinetic data were averaged.

Electrophoresis.

Denaturing PAGE analysis and oligomer purification were performed by using 16% slab gels 25 × 45 cm in 8 M urea/TBE at 1.5 kV for ≈ 2 h.

Sequencing.

To determine the depurination/cleavage position, the oligomer was sequenced by using a Sequenase 2.0 kit (USB Corp.) and ³²P-5′-end-labeled 8-nt primers complementary to the 3′ end of the 29-nt sickle cell β-globin gene fragment D-1.

Alkaline Phosphatase Treatment, Ligation, and 5′-End Labeling.

Standard protocols were followed.

Abbreviations

dG

deoxyguanosine

Exo III

exonuclease III.