Bypassing A Toxic DNA Adduct: A Look Into The Enzyme Active Site
By: Carol A. Rouzer, VICB Communications
Published: October 11, 2011
X-Ray crystallography reveals how a bypass polymerase overcomes the DNA damage caused by the highly carcinogenic mycotoxin aflatoxin B1.
Aflatoxin B1 (AFB1) is one of a number of mycotoxins produced by Aspergillus species of molds (Figure 1). AFB1 is highly carcinogenic, and mold contamination of stored grains, nuts, and seeds, particularly in warm humid climates, is believed to be the cause of high rates of hepatic cancer in some human populations. AFB1’s carcinogenicity results from its metabolism by cytochromes P450 to a highly reactive epoxide, AFB1-exo-8,9,-epoxide (Figure 2). This epoxide metabolite reacts with DNA, primarily at the N7 position of guanine bases, to produce the AFB1-N7-dG adduct. Further reaction of this adduct with hydroxide ion leads to opening of the five-membered ring of the guanine base to yield an AFB1-formamidopyrimidine (AFB1-FAPY) adduct (Figure 2). The effect of these two adducts on DNA replication has been studied extensively. Results suggest that both are mutagenic, leading to transversions from the guanine base to thymidine. The AFB1-FAPY adduct is more mutagenic than the AFB1-N7-dG adduct in bacteria. In addition, the AFB1-FAPY adduct persists in cells, suggesting that it is poorly repaired. The basis for these differences in mutagenicity between the AFB1-N7-dG and AFB1-FAPY adducts is poorly understood, leading Vanderbilt Institute of Chemical Biology (VICB) investigators Mike Stone and Martin Egli to examine the structures of these adducts in a model DNA polymerase enzyme [S. Banerjee et al. (2011) J. Amer. Chem. Soc., 133, 12556].
Figure 1. Electron micrograph of an asexual Aspergillus sp. fruiting body. Aspergillus is a common genus of fungus, a number of which produce AFB1. Of these, Aspergilus flavus is the major source of the toxin. Photomicrograph from the CDC (Robert Simmons provider, Janice Haney Carr photo credit) http://www.cdc.gov/nczved/divisions/dfbmd/diseases/aspergillosis/
Replicative DNA polymerases are primarily responsible for the highly accurate and rapid replication of DNA that precedes cell division. These enzymes are usually blocked by DNA adducts, however, so cells have a set of bypass polymerases specifically designed to deal with unrepaired damaged bases. In order to accomplish their goal, bypass polymerases have a larger, less restrictive active site that allows them to accommodate a bulky adduct, but this may lead to the incorporation of incorrect bases opposite the damage site. The result is that the cell can continue its replicative process, but it may do so with errors in the sequence of the newly synthesized DNA. Since replication in the presence of AFB1-DNA adducts occurs primarily through the action of bypass polymerases, the Stone and Egli labs used the bypass polymerase Dpo4 from the bacteria Sulfolobus solfataricus as a model system to study AFB1-dependent mutagenesis. They first used Dpo4 to replicate DNA template-primers containing the AFB1-N7-dG and AFB1-FAPY adducts in vitro to demonstrate that, indeed, the enzyme more efficiently bypassed the AFB1-N7-dG adduct than the AFB1-FAPY adduct. Furthermore, Dpo4 more incorrectly incorporated dA opposite the AFB1-FAPY adduct more frequently than opposite the AFB1-N7-dG adduct. This misincorporation, which would lead to guanine to thymine transversion mutations, is consistent with prior data obtained from in vivo mutagenesis studies in bacteria.
Figure 2. AFB1 is metabolized by a number of cytochromes P450, producing the reactive AFB1-exo-8,9-epoxide (top). The epoxide functional group is shown in red. The epoxide can then react at the N7 position of guanine bases in DNA to form the AFB1-N7-dG adduct. Reaction of this adduct with base (OH-, shown in blue) leads to opening of the five-membered ring of the guanine base, forming the AFB1-FAPY adduct.
To understand why Dpo4 more efficiently and more correctly bypasses AFB1-N7-dG than AFB1-FAPY, the Stone and Egli labs prepared crystals of the enzyme complexed with template-primers containing each adduct with or without an incoming nucleotide to be added to the primer. In all, they obtained diffraction quality crystals and solved the structures of six different complexes. The results showed that for both adducts, the structure of the adducted DNA bound in the Dpo4 active site was similar to that reported for free adducted DNA in solution. In both cases, the AFB1 planar ring system is inserted into the DNA helix on the 5′ side of the adducted base. However, because of the angle of the bond attaching AFB1 to dG in the AFB1-N7-dG adduct, the two ring systems are offset by an angle of 16o in that structure (Figure 3). In contrast, the opening of the five-membered ring of the adducted guanine in the AFB1-FAPY structure provides greater flexibility, allowing the AFB1 to be inserted into the DNA helix parallel to the adducted FAPY moiety (Figure 4). The wider angle in the case of the AFB1-N7-dG adduct provides a larger area for an incoming nucleotide to enter and align itself for addition to the nascent DNA chain. This observation provides a possible explanation for why this adduct is more efficiently bypassed than the AFB1-FAPY adduct.
Figure 3. Crystal structure of an AFB1-N7-dG adduct-containing template primer bound in the active site of Dpo4. An incoming dCTP forms Watson Crick hydrogen bonds with the adducted guanine base, placing it in position for addition to the primer. The AFB1 ring system of the adduct is intercalated into the DNA helix and lies at a 16o angle with respect to the guanine to which it is attached. The adduct is colored in cyan. Reproduced with permission from Banerjee et al. (2011) J. Amer. Chem. Soc., 133, 12556. Copyright American Chemical Society.
Figure 4. Crystal structure of an AFB1-FAPY adduct-containing template primer bound in the active site of Dpo4. The incorporated cytosine base opposite the adduct forms Watson Crick hydrogen bonds with the adducted FAPY six-membered ring. The AFB1ring system of the adduct is intercalated into the DNA helix and lies parallel to the FAPY moiety to which it is attached. The adduct is colored in red. Reproduced with permission from Banerjee et al. (2011) J. Amer. Chem. Soc., 133, 12556. Copyright American Chemical Society.
For both adducts, a cytosine base opposite the adducted guanine is able to align with conventional Watson Crick hydrogen bonding. Furthermore, for both adducts, an incoming dATP is able to pair to a thymine on the 5′ side of the adduct with conservation of conventional Watson Crick hydrogen bonds. However, in both cases, the distance between the phosphate group of the incoming dATP and the 3′-hydroxyl group to which it must bind is too large for reaction to take place easily. The investigators hypothesize that movement of the protein could allow these two groups to come close enough for bonding to occur. Also, for both adducts, an incoming dTTP is able to form hydrogen bonds to the thymine on the 5′ side of the adduct, but in this case, the misspaired base is positioned in such a way that addition to the nascent chain is unlikely. This finding is consistent with the in vitro replication data showing that addition of dATP is strongly preferred at this site.
The results provide the first structural insight into the process of AFB1 adduct bypass. They suggest a reason why the AFB1-N7-dG adduct is more readily and more accurately bypassed than the AFB1-FAPY adduct.. However, a general mechanism by which guanine to thymine transversions occur in the case of either adduct remains to be elucidated.