since our first biology class, we have learned that maintenance of DNA integrity is indispensable for the life of every mammalian cell. When DNA is damaged and the code cannot be properly deciphered, something significant invariably happens: mutation, evolution, disease (or at least “disease propensity”), senescence, etc. Not surprisingly, mammalian cells are endowed with complex and rather efficient surveillance mechanisms for detecting DNA damage and repairing such lesions with considerable precision. In what could be viewed as a final line of defense, changes in DNA integrity also trigger cell death pathways. Cells at risk for mutagenic events are thus eliminated before the organism suffers the consequences of having a constituent with a potentially defective genome. It seems to make good sense that the integrity of DNA is guarded at all costs.
Maybe not. There is mounting evidence that reversible alterations in genetic integrity may be a part of normal, day-to-day cellular operations. In a recent provocative report, Ju and colleagues (10) found in cultured MCF-7 cells that signal-dependent activation of gene transcription requires transient, topoisomerase IIβ (Topo IIβ)-mediated, site-specific formation of double strand (ds) breaks in DNA. In cultured pulmonary artery endothelial and smooth muscle cells, hypoxia and other signals reported to use reactive oxygen species (ROS) as second messengers have been shown to cause reversible oxidative base modifications in nuclear DNA (8, 16). Double strand DNA breaks and oxidative DNA “damage” associated with signaling! What for?
The dsDNA break discovered by Ju et al. (10) and findings of signaling-related oxidative base damage share a number of conspicuous features that may point to a common biological role. First, both kinds of “lesions” were detected in functionally significant regions of inducible genes. Using a clever adaptation of the ChIP assay, Ju et al. found dsDNA nicks in a hormone-responsive, nucleosome-associated promoter sequence of the pS2 gene. In terms of the VEGF gene hypoxic response element, we (8, 16) used ligation-mediated PCR to detect oxidative base modifications at nucleotide-resolution and found that the 3′-guanine of the hypoxia-inducible factor (HIF)-1 DNA binding sequence was most frequently targeted for oxidative modification in response to hypoxia and other stimuli using ROS as second messengers.
Second, multiple lines of evidence suggested that the lesions in both the pS2 and VEGF promoters somehow modulated gene expression. In both instances, changes in DNA integrity were temporally related to induction of gene expression. In the case of the pS2 promoter, Ju et al. (10) demonstrated that the Topo IIβ-mediated dsDNA break served as a substrate for poly[adenosine diphosphate(ADP)-ribose]polymerase-1 (PARP-1) binding, which in turn was necessary for local chromatin remodeling and transcriptional activation. In our studies, incorporation of a model abasic site at the hypoxia-modified guanine in an oligonucleotide sequence of the VEGF hypoxic response element was associated with increased HIF-1 binding and more robust hypoxia-induced reporter gene expression (16).
Another significant aspect of the Topo IIβ-mediated dsDNA break and the ROS-mediated DNA base modification products is that they identify a long-sought functional connection between initiation of transcription with sensing and repair of DNA lesions. In the case of pS2 promoter activation, Topo IIβ and PARP-1 are both well recognized for their involvement in the DNA damage and repair apparatus; in transcriptional activation, they play central roles by creating the dsDNA breaks, assembling other components of the transcriptional complex, and triggering localized loss of histone H1. With regard to VEGF expression, the bifunctional Ref-1/Ape1 is a key member of the hypoxia-inducible transcriptional complex (7, 15). The Ref-1 (redox effector factor) domain interacts with HIF-1 and other transcription factors and seems to be essential for their full transcriptional activity. The Ape1 domain, referring to apurinic-apyrimidinic endonuclease-1, functions in the second step of the base excision pathway repairing oxidative DNA damage (5).
Finally, both dsDNA breaks and base oxidation products cause prominent changes in the bendability and local sequence topology; this structural effect of DNA lesions also would be expected to impact fundamentally on transcriptional regulation. Promoter DNA is believed to wrap tightly around the histones comprising the nucleosome core particle, thus creating a steric obstacle to protein binding (6). During transcriptional activation, the promoter sequence must somehow become accessible to transcriptional proteins. Traditional concepts hold that chromatin remodeling factors are important for this process (6), but there is a chicken-and-egg situation inherent in this scenario: how do remodeling factors access the DNA and “relax” its association with the nucleosome if the DNA is wrapped so tightly that the remodeling factors themselves cannot gain access? Because short promoter sequences are believed to be relatively stiff (12, 14), it seems unlikely that transcriptional complexes, by themselves, impart sufficient energy to overcome the intrinsic rigidity of the DNA sequence, unwrapping it from the associated nucleosome. However, introduction of a dsDNA nick could serve to relax the DNA, while base oxidation products, such as 8-oxoguanine or its repair intermediate, an abasic site, introduce considerable tortional flexibility (1, 11). In this latter context, it is interesting that the hypoxia-modified guanine resides in the stiffest trinucleotide sequence of the VEGF hypoxic response element (16) where the incorporation of a “hinge” at this site in the form of a base oxidation product could be expected to have a dramatic effect on local sequence flexibility. Indeed, our preliminary observations indicate that incorporation of a single abasic site at the hypoxia-modified guanine engenders a remarkable degree of conformational flexibility upon nuclear protein binding (unpublished observations). The functional significance of these two different DNA structural changes, a dsDNA break and a base oxidation product, is thus similar; by changing mechanical properties of the promoter sequence, the modifications could enable DNA bending thereby reducing the steric obstruction to transcription factor binding.
These considerations suggest a new twist to current models of regulated gene expression (Fig. 1). With the use of DNA-cleaving enzymes or signaling-related ROS as tools, nature seems to dynamically impact the chemical integrity of local DNA sequences. The resulting change in DNA structure then facilitates or enables assembly of the transcriptional complex, including local chromatin remodeling factors, that serves to initiate gene expression.
Could it be true that complicated, multicellular organisms intentionally create promoter damage to facilitate normal gene regulation? The answer is presently unknown. In the case of signaling-related dsDNA breaks, the occurrence of such lesions in circumstances more physiological than cultured cells has not been reported. In terms of ROS-mediated base damage, biomarker or histopathological evidence for the common base oxidation product, 8-oxoguanine, has been observed in a variety of settings, including diabetes (3, 13), coronary artery disease (4), renovascular hypertension (9), and primary pulmonary hypertension (2), to name just a few. It is not apparent, however, whether the DNA damage seen in these disorders is a toxic response to oxidative stress or whether the base oxidation products are, at least in part, associated with signaling pathways using ROS as second messengers.
There is another, perhaps more philosophical, reason to consider that DNA damage might be a normal part of gene regulation. It is a cornerstone of evolution that all species have a finite lifespan. The bending or breaking of DNA that appears to be linked to gene expression must entail some risk to genetic integrity. If these chemical alterations in DNA are not properly repaired or accommodated, we are left with a situation wherein the signals required for the normal function of living cells might ultimately provide the stimulus for their dysfunction, senescence, and death.
Studies from the authors' laboratories were supported in part by National Heart, Lung, and Blood Institute Grants RO1-HL-058234, RO1-HL-073244, and PO1-HL-066299.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2007 the American Physiological Society