Tris-dependent oxidative DNA strand scission during electrophoresis

888 T. Ray, A. Mills and P. Dyson Electrophoresis 1995, 16, 888-894

Trevor Ray' Andrew Mills* Paul Dyson'

'Molecular Biology Research Group, School of Biological Sciences, 'Department of Chemistry, University of Wales Swansea, Singleton Park, Swansea

Tris-dependent oxidative DNA strand scission during electrophoresis The DNA of two Streptomyces species contains site-specific labile modifications. During gel electrophoresis the DNA can undergo Tris-dependent strand scission at the positions of these modifications. Our investigations into the nucleolytic activity which reacts with the modifications implicate a peracid derivative of Tris formed at the anode; the kinetics of production and decay of this activity were followed using both a DNA cleavage assay and a reduced methyl viologen assay to measure oxidant. Anode activation could be chemically mimicked by addition of peracetic acid to Tris buffers. We tested the DNA cleavage activity of several other compounds after anode or chemical activation; we used an analogue of Tris lacking a primary amine group and also several reagents known to promote DNA strand cleavage by amine-catalysis at abasic sites. Anode generation of oxidant could be detected for compounds containing either hydroxyl or carboxyl groups. However, DNA cleavage activity correlated with oxidant formation only for those compounds also containing primary amine groups. These results support a mechanism of DNA strand scission at modification sites via concerted peracid-mediated oxidative and amine-catalysed reactions. The novel finding of Tris-dependent formation of a long-lived reactive oxidant at the anode suggests that this compound is unsuited as an electrophoresis buffer for certain biological macromolecules.

1 Introduction

Conventional gel electrophoresis of nucleic acids, in par- ticular double-stranded DNA, requires an ionic buffer maintained at a pH of around 7.5. Historically and partly for economic reasons, Tris has been the buffer of choice for this purpose. The buffering capacity of Tris is dependent on its primary amine group. This group, however, can itself react with certain DNA lesions resulting in strand scission. Primary amines such as histidine, lysine, putrescine and Tris can cleave DNA containing abasic (AP) sites induced in vitro by either acid hydrol- ysis [l] or treatment with the antitumour antibiotics bleo- mycin and neocarzinostatin [2]. The structures of the AP sites induced by these three treatments are subtly different, but each possesses a free aldehyde group. Scis- sion of the DNA backbone is believed to involve (3-elimination through formation of a Schiff base linkage between the amine and the aldehyde of the AP site [3]. AP sites may occur spontaneously in DNA maintained at physiological pH, temperature and ionic strength [4]. Their rate of formation in vivo is low and, as the lesions are efficiently repaired, DNA isolated directly from an organism is generally stable when subsequently electro- phoresed in Tris buffers. In the case of nonsupercoiled covalently-closed or linear double-stranded DNA, cleavage at infrequent AP sites would in fact go unno- ticed. The complementary strand can be expected to be undamaged opposite the site of cleavage and the electro-

Correspondence: Dr. Paul Dyson, Molecular Biology Research Group, School of Biological Sciences, University of Wales Swansea, Singleton Park, Swansea SA2 IPP, Wales, UK (Tel: +44-1792-295667; Fax: +44- 1192-295441)

Abbreviations: AP, abasic (apurinic-apyrimidinic); HAE, HEPES- acetate EDTA buffer pH 7.5; MeT, Tris(hydroxymethy1)ethane; TAE, Tris-acetate EDTA buffer pH 7.5

Keywords: Streptomyces / DNA modification / Oxidative DNA strand scission

phoretic mobility of a molecule containing non-opposed single-strand breaks is indistinguishable from that of the corresponding intact double-stranded molecule.

We have been investigating an unusual type of naturally occurring site-specific DNA modification. Modified DNA undergoes Tris-dependent site-specific double-strand cleavage during electrophoresis IS]. The modifications have been identified in the DNA of two unrelated gram- positive antibiotic-producing Streptomyces species: s. lividuns [6, 71, the member of the genus most widely used for genetic manipulation, and S. avermitilis [8], a pro- ducer of commercially important antihelminthic sec- ondary metabolites called avermectins. The modifications exhibit an identical site-specificity in both species, and are located closely-opposed on both strands within a G+C rich sequence [9]. The structure of the modification is unknown, although the sugar phosphate backbone of the DNA remains intact, permitting isolation of modified supercoiled plasmid molecules. In vitro reactivity during electrophoresis results in partial double-strand cleavage and there is significant variance in the amount of cleavage observed at different modification sites. This could be a consequence of the DNA being only partially modified in vivo with the extent of modification varying between different sites. The modified base is estimated to constitute 4 0.1% of the total base composition.

In contrast to the reaction of Tris with AP sites, no cleavage is observed if modified DNA is incubated in Tris buffer in the absence of electrophoresis [ 5 ] . The nucleolytic species which reacts with the modifications is rapidly generated at the anode during electrophoresis and has a half-life of around 15 hours. The involvement of Tris was demonstrated by adding this compound to a nonreactive alternative electrophoresis buffer based on HEPES. This buffer had previously been developed to permit stable electrophoresis of DNA containing bleo- mycin or neocarzinostatin-induced AP sites [2]. Full nucleolytic activity could be generated at the anode in

0 VCH Verlagsgesellschaft mbH, 69451 Weinheim, 1995 0173-0835/95/0606-0888 $5.00+.25/0

Electrophoresis 1995, 16, 888-894 Tris-dependent oxidative DNA strand scission 889

HEPES buffers supplemented with at least 10 mM Tris. Electrochemical reactions at the anode can be expected to generate oxidants, some of which may be free-radical species. The involvement of such species in buffer activation was demonstrated by inhibition with micromolar concentrations of thiourea [5] , which is both a radical scavenger and reducing agent. Stable electrophoresis of the modified DNA can therefore be achieved by using a non-primary amine buffer such as HEPES or by adding thiourea to Tris buffers [lo].

We have now undertaken further characterization of the nucleolytic species. Supporting evidence from inhibition studies, direct measurement of the oxidant and chemical mimicry of anode activation together suggest formation of a peracid derivative of Tris. DNA cleavage is by a bifunctional agent: the presumptive peracid group is implicated and, as with cleavage at AP sites, we demonstrate that the primary amine group is intimately involved in reactions with the site-specific DNA modifications. The evidence for Tris-dependent formation of a reactive oxidant at the anode demonstrates that this buffer is far from optimal for the analytical electrophoresis of biological macromolecules.

2 Materials and methods

2.1 General DNA manipulation

Plasmid pIJ699 [Il l DNA isolated from S. lividans strain 66 was used throughout these studies. The plasmid was purified following alkaline lysis [12] as previously described [ 5 ] . The DNA was linearised with EcoRI (Gibco BRL) prior to induction of site-specific cleavage. Exonuc- lease 111 (Amersham) digestions were carried out for 20 min at 37°C in a buffer recommended for AP endonuclease activity [13], containing 100 mM Tris-HCl, pH 8.0, 50 mM MgCl,, 50 mM 6-mercaptoethanol, 50 pg/mL BSA. Reacted DNA samples were analysed for site- specific cleavage following gel electrophoresis in an HEPES buffer (HAE, see below) and subsequent Southern hybridization as previously described [5]. Prep- aration of random-primed digoxigenin-labelled DNA probes was performed according to the instructions pro- vided by the supplier (Boehringer, Mannheim). To quan- tify DNA cleavage, individual lanes of a blot were scanned using a Joyce-Loebl densitometer. DNA cleavage reactions were repeated a minimum of three times; tabulated and plotted results represent mean values obtained (standard deviations were less than or equal to 20% of these values).

2.2 Preparation of test buffers for DNA cleavage

Chemicals were obtained from Sigma or Aldrich. The standard Tris-acetate-EDTA (TAE) buffer contained 40 mM Tris, 20 mM sodium acetate, 0.8 mM EDTA, pH 7.5, adjusted with acetic acid. The standard HEPES-acetate- EDTA (HAE) buffer contained equivalent molarities with HEPES substituting for Tris; the final pH was adjusted to 7.5 by addition of sodium hydroxide. The activity of Tris(hydroxymethy1)ethane [MeT] (source: Aldrich) was tested by addition of 40 mM MeT to HAE.

A 250 mL sample of a buffer, maintained at 37"C, was electrophoretically activated for DNA cleavage by placing it in a horizontal submarine gel chamber (Bio-Rad) with platinum electrodes, and applying a constant voltage of 80 V. After a given time for activation, normally 10 min, 500 pL of the buffer was sampled from approximately 5 mm adjacent to the anode and added to 0.5 pg DNA in 5 pL H,O in a microfuge tube. The cleavage reaction was then allowed to proceed for 2 h at 37 "C. The reaction was terminated by addition of 50 pL 3 M sodium acetate, pH 5 , and 550 pL isopropanol. DNA was precipitated at -2O"C, prior to centrifugation, drying the pellet, resuspension and gel loading. To study inhibition, a pre- warmed 450 pL buffer sample was added to a 50 pL solution of inhibitor (10 X final concentration) prepared in H,O (with the exception of thiophenol, prepared in 40% ethanol). After brief mixing, the buffer-inhibitor solution was incubated at 37°C for 15 min prior to combining it with the DNA. For chemical activation, a 450 pL buffer sample maintained at 37°C was combined with 50 pL of a peracetic acid solution (10 X final concentration). After brief mixing, the solution was combined together with 0.5 pg DNA and incubated for 2 h at 37°C. Reactions were terminated and the DNA recovered by isopropanol precipitation. Cleavage at AP-sites was assayed by incu- bating treated DNA in 500 pL volumes of given concentrations of AP-site reactive reagents for 2 h at 37 "C, prior to recovery of the DNA by precipitation.

2.3 The reduced methyl viologen assay

Reduced methyl viologen (MV.') is a strong reducing agent (EO' = -0.44 V) and has an intense blue colour; it can be readily reoxidised to the colourless MVZ+, this oxidation being spectrophotometrically detectable at 605 nm [14]. Three mL of a 2 mM methyl viologen (MV") dichloride (Aldrich) solution in HAE was added to a glass cuvette containing a magnetic stirring bar. A septum was fitted, through which two long-form Pasteur pipettes passed, one to allow application of nitrogen and the other to vent the gas. A zinc rod, used to reduce the methyl viologen solution, also pierced the septum. The head space above the solution in the cuvette was purged for 2 min with oxygen-free nitrogen. The gas inlet pipette was then lowered into the solution along with the zinc rod and the magnetic stirrer switched on. Reduc- tion of the methyl viologen solution was monitored on a microcomputer and allowed to proceed until an absorbance reading of approximately 2 units was obtained (equivalent to 175 p~ MV.'). At this point both the gas inlet and zinc rod were lifted out of solution. Samples to be assayed were placed in a glass bottle, sealed with a septum, and purged with nitrogen for 90 s. Samples were removed using a 1 mL syringe and hypodermic needle and 0.3 mL was injected into the cuvette containing MV.'. The change in absorbance was recorded on a microcomputer. A compensation factor for change in absorbance due to dilution of MV.' was included when assaying the concentration of oxidant in test buffers. Oxidant concentration was calculated using the equa- tion:

A = Eel (1)


where A is A absorbance at 605 nm, E the extinction coefficient for MV' (11400 M-' cm-' [15]), c the concentration (M), and I the cell path length (1 cm). Values plotted represent mean vaIues obtained from three sep- arate experiments (variance between experiments was within +/- 1 standard deviations of plotted values).

3 Results

3.1 Inhibition of DNA cleavage by reducing agents

The ability of thiourea to inhibit DNA cleavage could be attributed to its activity as either a radical scavenger or a reducing agent. To discriminate between these possibili- ties, inhibition was tested with a number of free-radical scavengers [16]: dimethylsulphoxide and mannitol (specific for hydroxyl radicals, *OH) and ethanol (specific for alkoxyl radicals, ROO). No inhibition was observed with concentrations of up to 50 mM of each scavenger (data not shown). This contrasts with the micromolar concentration of thiourea required for inhibition of DNA cleavage, suggesting that the nucleolytic species was a nonradical oxidant. Competition reactions with reducing agents of known redox potential were then carried out. Table 1 lists the reducing agents tested, their respective E,' values, configuration of sulphur groups and dissociation constants of thiol groups where relevant, and the concentrations required for complete inhibition of DNA cleavage. These results indicated that the reducing agents fell into two classes of compound. The most effective at inhibiting DNA cleavage contained incomple- tely oxidised sulphur moieties. Nonsulphur-containing compounds were less effective at cleavage inhibition. A correlation between redox potential and the ability to inhibit DNA cleavage was observed for this second class of compounds. For the compounds containing incomple- tely oxidised sulphur groups, inhibition appeared to be dependent also on the configuration of the sulphur and, for those containing thiols, the dissociation constant of these groups.

3.2 Measurement of the oxidant using reduced methyl viologen

A sensitive spectrophotometric assay to measure production of the oxidant was employed using the redox indi-

Table 1. Reducing agents tested Sulphur Concentration

Compound EO' group p K s ~ for complete

Sodium dithionite -0.53 s20:- N/Aa) 5 W Thiourea -0.48 R-CS-R NIA 5 CLM Thiophenol -0.30 R-SH 7.6 5 W Glutathione -0.24 R-SH 8.6 5 W DTT -0.33 HS-R-SH 9.2 50 NM Cysteine -0.22 R-SH 8.7 50 WM Sodium sulphite -0.45 s 0 3 2 - NIA 50 PM

NADH -0.32 N /A NIA 500 FM Isopropanol -0.29 NIA NIA No inhibition Riboflavin -0.21 NIA NIA No inhibition Ethanol -0.20 NIA N/A No inhibition

a) N/A, not applicable

configuration inhibition

Thioglycolate -0.14 R-SH 10.4 500 FM

cator methyl viologen (l-l'-dimethyl-4,4'-bypyridiium dichloride). Formation of the oxidant at the anode was investigated as a function of time after switching on the power-pack. Prior to passing a current, no oxidant could be detected with either TAE or HAE buffers. We were also unable to measure an oxidant when a current was passed through HAE buffer for up to 30 min. In contrast, when TAE was electrophoretically activated, an oxidant was rapidly formed. Anode samples were tested in parallel for their DNA cleavage activity and presence of oxidant (Fig. 1). Maximal DNA cleavage was observed after 4 min of anode activation, representing a concentration of 150 PM oxidant formed in the buffer itself (AA = 0.164, equivalent to 14.5 WM oxidant in 3.3 mL of diluted buffer). Although the concentration of oxidant continued to rise, a deviation from a linear increase was noted after this time. Following 30 min activation, there was 290 WM oxidant present in anode samples. No significant further increase in oxidant concentration was detectable after this time point (activation was continued until 60 min). A second series of experiments was performed whereby decay of the oxidant was monitored. TAE was activated for 30 min and samples were tested at specific times after the power-pack had been switched

- - 30.3 A-

100 I

LJ

u 40 ............................................................

................................................................ n 11 4 0.1 . . * r

.................................................................. I

1 0'05 0.0

0 5 10 15 20 25 30 Time (mins)

-m- -&- DNA Cleavage Methyl Viologen

Figure 1. Correlation between anode generation of both oxidising species and the DNA cleavage activity. TAE was sampled from adjacent to the anode at time points after switching on the power pack (constant voltage, 80 V; resultant current, 150 mA). Oxidant was measured as the change in absorbance at 605 nm using a reduced methyl viologen assay. DNA cleavage was quantified by determining the relative area of a 5.7 kbp cleavage product by densitometry; cleavage activity is expressed as a percentage of the value observed at the 30 min time point.

Electrophoresis 1995, 16. 888-894 Tris-dependent oxidative DNA strand scission 891

..............................................

0.25

cb Y d

0.1 ti

t 0 ' I I I ' 0

Time (h)

+ -&-

0 10 20 30

DNA Cleavage Methyl Viologen

Figure 2. Correlation between decay of both the oxidising species and the DNA cleavage activity. TAE was anode-activated for 30 min and, at time points over a 35 h period, the remaining DNA cleavage activity and oxidant present were quantified as in Fig. 1.

off. Both DNA cleavage activity and the presence of oxidant were assayed (Fig. 2). Decay of both activities correlated.

3.3 Chemical activation of Tris buffer

Whilst the inhibition studies suggested that free radicals were not directly involved in DNA cleavage, dimeriza- tion of anode-generated hydroxyl radicals could have formed small amounts of hydrogen peroxide. To deter- mine if this was the oxidant responsible for DNA cleavage, inactive TAE and HAE were supplemented with between 10 PM and 10 mM H202. No DNA cleavage was observed for either buffer with addition of H,02 (data not shown). This suggested that neither H,O, nor its breakdown products could be involved in DNA cleavage. During electrophoresis, levels of H,O, generated at the anode could be expected to be similar for either TAE or HAE; that no oxidant could be measured in anode-sampled HAE using the methyl viologen assay suggested that H,O, generation is minimal and that a quite different species must be generated in Tris buffers.

We speculated that a peracid derivative of Tris could be formed by oxidation of hydroxymethyl groups at the anode, following the reaction scheme:

0 0 II

RCOOH 'I RCZO RCH,OH '2 RCHO 4 RC.0 2 RCOO

To test if peracid groups could be implicated in DNA cleavage, peracetic acid, at concentrations between 1 J ~ M and 10 mM, was added to both inactive TAE and HAE. When peracetic acid was added to inactive HAE there was no induction of DNA cleavage. When added to inactive TAE, however, it caused a significant level of cleavage at concentrations between 100 VM and 10 mM (Fig. 3A). Maximal cleavage was obtained with 500 VM peracetic acid; a reduction of cleavage was observed at higher concentrations, possibly due to instability of the nucleolytic species in strongly oxidising conditions. The range of concentrations of peracetic acid used was effective at complete reoxidation of reduced methyl viologen. As the redox potential of the oxidant present in anode- activated samples of TAE was considerably lower, we suspected that peracetic acid was not reacting directly with the DNA modifications to cause strand scission. This was supported by the lack of DNA cleavage in HAE supplemented with peracetic acid.

We considered two possible explanations: (i) peracetic acid was oxidising one or more hydroxymethyl groups of Tris and the product thus formed could subsequently mediate DNA cleavage, or (ii) peracetic acid reacted with the DNA modifications to form cleavage intermediates, for example AP sites, susceptible to amine attack. To test the latter, modified DNA was incubated with HAE supplemented with peracetic acid and, following precipitation, the recovered nucleic acid was incubated with inactive TAE or other AP-site reactive reagents and then examined for site-specific cleavage. The reagents tested were 40 mM and 100 mM putrescine, 40 mM histidine, 40 mM lysine, 0.2 M KH,PO,:K,HPO,, pH 11.5 (most AP sites are alkali labile [17, 18]), and 5 units of the AP endonuclease, exonuclease 111. None of the treatments resulted in DNA cleavage (data not shown), indicating that AP sites were not formed by reaction of peracetic acid with the DNA modifications.

3.4 The primary amine group of Tris is involved in DNA

To investigate the role of the amine group in DNA cleavage, we employed an analogue of Tris, Tris(hydroxy- methy1)ethane [Methyl Tris; MeT], in which the primary amine group is replaced by a methyl group. As this compound has no buffering capacity, electrophoresis buffers were prepared consisting of HAE containing 40 mM MeT (MeTHAE). DNA incubated with anode-activated MeTHAE was not cleaved, suggesting that the amine group of Tris is required for strand scission (Fig. 3B). Using the methyl viologen assay, however, we could measure up to 500 PM oxidant present after anode activation of MeTHAE. Thus it was possible that the oxidant could react with the DNA modifications to form cleavage intermediates susceptible to amine attack. DNA incubated with activated MeTHAE was precipitated and the recovered nucleic acid subsequently treated with either inactive TAE, 40 mM putrescine, 0.2 M KH,PO,:K,- HPO,, pH 11.5, or exonuclease 111. None of these treatments resulted in DNA cleavage (data not shown), indicating that stable intermediates or AP sites were not formed by reaction with activated MeTHAE. These results suggested that DNA cleavage by the nucleolytic

cleavage


A B

C

Figure 3. Site-specific cleavage of S. lividuns pIJ699 DNA. Linearised pIJ69Y DNA was incubated for 2 h at 37OC with various test buffers: (A) TAE (lane l), anode-activated TAE (2), TAE with 10 mM peracetic acid (3), TAE with 1 mM peracetic acid (4), TAE with 500 PM peracetic acid (S), TAE with 100 WM peracetic acid (6), TAE with 10 I.IM peracetic acid (7). (B) MeTHAE (lane l), anode-activated MeTHAE (2), 40 mM Tris combined with anode-activated MeTHAE (3), 40 mM putrescine combined with anode-activated MeTHAE (4), 100 mM putrescine combined with anode-activated MeTHAE (5). (C) Histidine-AE (lane l), lysine-AE (2), anode activated histidine-AE (3), anode-activated lysine-AE (4), putrescine-AE (5 ) , anode-activated putrescine-AE (6), HAE with 1 mM peracetic acid (7), TAE with 1 mM peracetic acid (S), histidine-AE with 1 mM peracetic acid (Y), lysine-AE with 1 mM peracetic acid (lo), putrescine-AE with 1 mM peracetic acid (11).

component of Tris buffers could be via a concerted reaction involving both the primary amine and presumptive peracid groups. To investigate this possibility, MeTHAE before and after anode activation was combined with DNA in the presence of 40 mM Tris or 40 mM and 100 mM putrescine (Fig. 3B). When inactive MeTHAE was added to DNA with Tris or putrescine there was no induction of cleavage. When anode-activated MeTHAE was added, cleavage was observed in the presence of Tris but not putrescine.

3.5 DNA cleavage by other bifunctional reagents

Lysine and histidine are two AP-site reactive reagents which, as well as possessing primary amine groups, could potentially undergo oxidation at their respective carboxyl groups to generate peracid derivatives. In this re- spect they differ from putrescine, which lacks oxygen- containing groups. We tested all three compounds in parallel for DNA cleavage activity. Equivalent solutions to TAE were prepared containing the test compound (40

Electrophoresis 1995, 16, 888-894 Tris-dependent oxidative DNA strand scission 893

mM> substituting for Tris, and adjusted to pH 7.5. Modi- fied DNA was incubated with samples prior to and after electrophoretic or chemical (solutions containing 1 mM or 10 mM peracetic acid) activation. Electrophoretically activated samples were also tested for the presence of oxidant using the methyl viologen assay. DNA cleavage was observed for electrophoretically or chemically activated histidine or lysine solutions (Fig. 3C). Prior to activation, histidine or lysine solutions, as well as putrescine solutions tested before and after activation, exhibited no DNA cleavage activity. The generation of nucleolytic species in electrophoretically activated histidine or lysine solutions correlated with formation of, respectively, 160 VM and 210 WM oxidant in these samples; no oxidant was detected in activated putrescine solutions.

4 Discussion

The nucleolytic activity present in anode-activated TAE and specific for modified Streptomyces DNA could be inhibited by addition of several reducing agents. Inhibi- tion by non-sulphur or fully-oxidised sulphur-containing reductants implicated generation of between 50 to 500 p~ of a mild oxidant at the anode. This correlated with the concentration of oxidant in activated TAE solutions measurable using a reduced methyl viologen assay. The kinetics of both production and decay of this oxidant and the nucleolytic activity were similar, confirming that they were one and the same. A plausible activation scheme suggested formation of a peracid derivative of Tris at the anode. This was supported by the finding that Tris could be chemically activated by peracetic acid to generate the nucleolytic activity. Peracetic acid itself could not cleave DNA, indicating that peracid-mediated oxidation of the DNA modifications was insufficient in itself to mediate strand cleavage. Likewise, anode activation of a nonprimary amine analogue of Tris resulted in formation of a presumptive peracid derivative but no DNA cleavage activity.

Amine-catalysed polynucleotide chain scission via p-elimination at AP sites is well documented. As Tris is known as an AP-site-reactive reagent, we suspected involvement of the NH, group in strand cleavage at the sites of the Streptomyces DNA modifications. We could find no evidence, however, for AP sites being stable reaction intermediates after peracid-mediated oxidation of the DNA. For example, solutions containing both putrescine, a good AP-site-reactive reagent, and nonprimary amine peracids could not cause site-specific DNA cleavage. Moreover, from these observation we could dis- count a two-step cleavage mechanism involving initial nucleophilic attack by an amine group followed by oxidation by a peracid group.

We suspected that strand cleavage could be via a concerted reaction between the DNA modification and a bifunctional reagent containing both peracid and primary amine groups. This was corroborated by the activity after anode or chemical activation of two other AP-site reactive reagents, histidine and lysine, which con- tain, in addition to their primary amine groups, carboxyl groups which could undergo oxidation to form peracid

groups. Tris-mediated amine-catalysed reactions with free aldehyde groups present in macromolecules are precedented. Our investigations have revealed a novel activity associated with Tris particularly when it is used in electrophoresis. When a current in passed through a Tris buffer, micromolar amounts of a relatively long-lived oxidant, presumed to be a peracid, are generated at the anode. Although this was revealed by reaction of the oxidant with a specific DNA modification, whose structure we are currently investigating, we believe that other components of macromolecules could also be susceptible to similar oxidation reactions. Oxidations mediated by peracids are known to be relatively nonspecific and to proceed rapidly under mild reaction conditions [19, 201. Indeed, electrophilic attack by various peracids on the normal nucleotides present in DNA can lead to N-oxidation of the bases [211. This process is mild enough to pre- clude adventitious polynucleotide strand scission, and occurs in neutral aqueous solutions at ambient temperature. N-oxidation of bases may well occur during electrophoresis of DNA in Tris buffers; it would not have been detected in this study.

At the outset of these investigations, we were interested in learning more about the reactivity of the site-specific modifications present in Streptomyces DNA. The finding that a peracid derivative of Tris is involved, however, has wider implications. Our results demonstrate that Tris is inappropriate as a biological buffer for optimal analytical electrophoresis of certain macromolecules. Unwanted and potentially destructive reactions involving oxidation and/or amine-catalysis can be avoided by substitution with alternative buffering compounds such as HEPES.

These investigations were supported by grants from the Society .for General Microbiology and the Nuf3eld Founda- tion. T R. was supported by an SERC studentship.

Received December 12, 1994

5 References

[ l ] Lindahl, T., Andersson, A,, Biochemistry 1972, I I , 3618-3623. [2] Porvirk, L. F., Houlgrave, C. W., Biochemistry 1988, 27, 3850-3857. [3] Coombs, M. M., Livingston, D. C., Biochim. Biophys. Acta 1968,

[4] Lindahl, T., Nyberg, B., Biochemistry 1972, 11 , 3610-3618. [5] Ray, T., Weaden, J., Dyson, P., FEMS Microbiol. Lett. 1992, 96,

[6] Dyson, P., Betzler, M., Kumar, T., Schrernpf, H., in: Alacevic, M., Hranueli, D., Toman, D. (Eds.), Genetics of Industrial Microorga- nisms, Proceedings of the Fifth International Symposium, Part B. Pliva, Zagreb 1986, pp. 57-65.

[7] Zhou, X., Deng, Z., Firmin, J. L., Hopwood, D. A., Kieser, T., Nucleic Acids Res. 1988, 16, 4341-4352.

181 Evans, M., Kaczmarek, F. S., Stutzman-Engwall, K., Dyson, P., Microbiology 1994, 140, 1367-1371.

191 Dyson, P. J., Schrempf, H., Heredity 1988, 61, 284.

174, 161-173.

247-252.

[lo] Evans, M., Dyson, P., Trends Genet. 1993, 9, 72. [ l l ] Kieser, T., Melton, R. E., Gene 1988, 65, 83-91. [12] Sambrook, J., Fritsch, E. F., Maniatis, T., Molecular Cloning:

A Laboratory Manual, Cold Spring Harbor Laboratory Press, New York 1989.

[13] Rogers, S . G., Weiss, B., in: Grossman, L., Moldave, K. (Eds.), Methods in Enzymology, Volume 65, Nucleic Acids (Part I), Aca- demic Press, San Diego 1980, pp. 201-211.

894 T. Ray, A. Mills and P. Dyson Electrophoresis 1995. 16, 888-894

[14] Summers L. A,, The Bipyridinium Herbicides, Academic Press,

[15] Mills, A., Douglas, P., Russell, T., J. Chem. Soc., Faraday Trans.

[16] Halliwell, B., Gutteridge, J. M. C., Free Radicals in Biology and

[17] Kochetkov, N. V., Budovski, E. I., Organic Chemistry of Nucleic

London 1980.

1990, 86, 1417-1423.

Medicine, Clarendon Press, Oxford 1989.

Acids, Plenum, New York 1972, pp. 425-448.

[18] Burger, R. M., Peisach, J., Honvitz, S. B., J . B i d . Chem. 1982, 257,

[19] Swern, D., Chem. Rev. 1948, 45, 1-68. [20] Lawesson, S.-O., Schroll, G., in: Patai, S. (Ed.), The Chemistry of

Carboxylic Acids and Esters, John Wiley & Sons, Chichester, UK

[21] Brown, D. M., in: Ts’o, P. 0. P. (Ed.), Basic Principles in Nucleic Acid Chemistry, Volume II, Academic Press, New York 1974,

3372-337s.

1969, pp. 669-703.

pp. 1-90.

Documents

Tris-dependent oxidative DNA strand scission during electrophoresis