A New Trick (Hydroxyl Radical Generation) For An Old Vitamin (B₁₂)

Abstract

Photolysis of hydroxocobalamin in the presence of plasmid DNA (pBR322) results in DNA cleavage. Temporal control of hydroxyl radical production and DNA strand scission by hydroxocobalamin was demonstrated using a 2-deoxyribose assay and a plasmid relaxation assay, respectively. The light-driven hydroxocobalamin-mediated catalytic formation of hydroxyl radicals was demonstrated using radical scavenging studies of DNA cleavage and via recycling of a hydroxocobalamin-resin several times without loss of efficacy.

Hydroxyl radicals (·OH) have been used to assess DNA structure¹ and to elucidate DNA-binding sites for proteins² and small molecules.³ An example of the Fenton reaction, in which Fe(II)·EDTA reduces H₂O₂, is the most commonly used method to generate ·OH. However, for intracellular purposes, a limitation of this method is the inability to precisely control when the reaction starts and stops. A Fenton-like reaction utilizing a caged Cu(II) ion that can be released by photolysis of the ligand cage has been reported.⁴ This strategy furnishes control over the initiation of the Fenton reaction, albeit not the termination step. Several light-mediated “Fenton reagents” have been developed.⁵ In contrast to the Fenton reaction, the majority of these compounds have the disadvantage of stoichiometric radical production. TiO₂ does catalytically produce ·OH upon exposure to light,⁶ however since TiO₂ induces intracellular oxidative damage in the absence of light,⁷ it cannot be used in a light-dependent fashion in cell-based studies. By contrast, synchrotron X-ray⁸ and γ⁹ irradiation can be used to precisely initiate and terminate ·OH production. Although these methods are advantageous for their ability to provide time resolved information, they are not widely available and they cannot be used to direct oxidative damage to specific intracellular or molecular sites.

Cobalamins, a class of compounds of which Vitamin B₁₂ is a member, could be useful for generating radical species under cellular conditions. For example, photolysis of the cobalt-carbon bond in methylcobalamin (MeCbl; 1 where X = Me) generates cob(II)alamin (B_12R) and methyl radical (·CH₃) (Scheme 1A).¹⁰ Based on this precedent, we wondered whether hydroxocobalamin (B_12a where X = OH) could undergo an analogous reaction to generate ·OH and B_12R (Scheme 1B). In the presence of oxygen, the latter species is rapidly oxidized to regenerate B_12a.¹⁰ We therefore investigated the possibility that the combination of light and B_12a could be used to oxidatively damage biomolecules via the generation of ·OH.

Scheme 1.

(A) Photolysis of MeCbl to furnish CH3· and (B) proposed light-dependent generation of ·OH by B_12a.

Initial studies employed a plasmid relaxation assay to observe the ·OH-mediated conversion of circular, supercoiled DNA (Form I) to relaxed, circular DNA (Form II). A Pyrex filter (>300 nm) was used to photolyze B_12a in the presence of pBR322 DNA. Analysis by agarose gel electrophoresis revealed a band corresponding to Form II DNA, but only from the light-exposed reaction mixtures containing B_12a (Fig. S1). By contrast, light in the absence of B_12a or B_12a in the absence of light, does not result in significant DNA damage.

DNA cleavage by B_12a does not occur in the absence of O₂, an observation consistent with a radical-mediated mechanism of DNA damage (Fig. S3).¹¹ In addition, the ·OH scavengers sorbitol¹² and sodium benzoate,¹³ efficiently prevent DNA cleavage (Fig. S5), thereby supporting the notion that damage is mediated by ·OH. The light-dependence of DNA cleavage was assessed by photolyzing B_12a for various time periods followed by incubation in the dark (Fig. S8). Short photolysis periods followed by dark incubation resulted in significantly less strand scission than photolyzing for the same total time period, demonstrating that the species formed by the photolysis of B_12a does not continue to damage DNA without continued exposure to light.

We employed an established ·OH assay to assess light-mediated production of ·OH.¹⁴ Deoxyribose is attacked by ·OH, generating malondialdehyde that, upon heating with thiobarbituric acid (TBA), produces a chromophore (532 nm).¹⁵ Unfortunately, we found that the presence of B_12a during the heating step interferes with the assay. Therefore, B_12a was immobilized on an aminebearing resin (2) using 1,1′-carbonyldi-(1,2,4-triazole) (Fig. S10). Deoxyribose was incubated with the B_12a-resin under photolytic and non-photolytic conditions and the resin was subsequently removed prior to the heating step. The light-driven nature of B_12a-resin-mediated ·OH production is evident by the time-dependent production of the TBA-malondialdehyde chromophore in the presence (Fig. 1, average absorbance increase per period: 0.104), but not in the absence (average absorbance increase per period: 0.003), of light.

Figure 1.

Light-driven ·OH production via photolysis of the B_12a-resin conjugate 2. The conjugate 2 was suspended in a deoxyribose solution (50 mM in pH 7.4 PBS). The suspension was sequentially photolyzed (60 min total), stored in the dark (100 min total), and photolyzed (60 min total), with aliquots removed and analyzed at various time points.

Photolysis of the Co-methyl bond in MeCbl generates Co(II)-cobalamin and ·CH₃ (Scheme 1A). The former, upon oxidation, produces B_12a,¹⁰ which should generate DNA damaging ·OH under photolytic conditions (Scheme 1B). However, ·CH₃ is also capable of causing DNA strand scission.¹⁶ We assessed the relative contribution of these two radical species to DNA damage using sorbitol and sodium benzoate (·OH traps) and 2,2,6,6-tetramethylpiperidine-1-oxyl (TEMPO; a carbon-centered radical trap¹⁷). TEMPO does not have an appreciable effect on the ability of MeCbl to damage DNA, but sorbitol and sodium benzoate almost completely protect DNA from damage (Fig. S11). Therefore, the initial stoichiometric production of ·CH₃ is small compared to that of the B_12a-mediated generation of ·OH.

The notion that B_12a catalytically produces ·OH is supported by radical trapping studies of DNA damage by re-using the B_12a-resin 2 several times without loss in its capacity to damage deoxyribose (Fig. S13). Furthermore, a B_12a concentration less than the pBR322 concentration was used to demonstrate catalytic production of strand scission events. Based on the number of observable strand scission events, B_12a induced a minimum of 20 damaging events per molecule during 90 min of light exposure (Fig. 2); an underestimate since plasmid relaxation assays do not visualize all DNA nicks. The amount of DNA damage per molecule increased with longer exposure to light. A band corresponding to linear DNA (form III) resulting from nonrandom double-strand cleavage was observed (Figs. S15-16),¹⁸ indicative of two radicals originating near each other in time and in space.

Figure 2.

Light-mediated DNA (30 μM/bp in pH 7.4 PBS) by hydroxocobalamin (300 pM). Samples were irradiated with Pyrex-filtered light from a mercury arc lamp.

A B_12a-spermine conjugate 3 was synthesized (Fig. S17) and proved useful for directing damage to DNA in an environment with a high concentration of a hydroxyl radical trap. Conjugate 3 forms a DNA complex with an apparent binding constant (K_app = 1.5 × 10⁸ ± 3.4 × 10⁶ M^-1) similar to the reported value for spermine (5.9 × 10⁸ M^-1).¹⁹ The B_12a-spermine conjugate induces DNA damage more effectively than B_12a (Fig. S20). More importantly, 10000 equiv. of sorbitol failed to significantly inhibit DNA damage induced by 3, whereas only 100 equiv. of sorbitol significantly inhibited strand scission due to B_12a (Fig. S22), suggesting that 3 generates ·OH near the DNA 3 it is bound to.

Light-dependent ·OH production by B_12a enjoys several characteristics that make it potentially useful as a generator of ·OH for biological studies. Light is increasingly employed in modern cell biology to control intracellular processes via cell-embedded photosensitive bioactive reagents.²⁰ The combination of light and light-sensitive reagents provides the investigator with exquisite temporal control over cellular biochemistry. Spatial control of oxidative damage is afforded by the addition of a binding moiety to B_12a because the highly reactive ·OH is generated proximal to the target biomolecule. In addition, light-induced radical generation with B_12a is catalytic, whereas radical generation by other systems is stoichiometric. Future research will focus on conjugating recognition moieties to B_12a to direct oxidative damage to specific macromolecules and organelles.