The DNA in our cells provides the blue-print of our body temporally as well as spatially. Evolution has placed a high premium on keeping this beautiful molecule stable throughout the lifetime of the organism, with DNA instability leading to mutations and genomic alterations, and possibly oncogenesis. Standing in almost direct opposition to this dogma, our laboratory is studying a group of enzymes called DNA deaminases, which induce DNA instability via deaminating dC residues to dU. Originally isolated as necessary factors for the development of our immune system, they have now been implicated in a number of fundamental aspects of cellular and organismal physiology, thereby providing the outline of our working hypothesis: DNA instability is an evolutionary desirable phenomena and will help in survival of the organism.
Figure 1. Schematic overview of the regulation and function of DNA deaminases that are being studied in our laboratory.
We are working on how AID is Activated in terms of transcriptional regulation, and how AID is Targeted and gains access to the specific loci. At the ssDNA target site, we are interested in what types of enzymatic Kinetics are regulating the activity. Once the deaminated dU (or dT) is present in the reformed dsDNA as a DNA mismatch, we are studying how this lesion is processed to induce immune diversity in Acquired Immunity and protects our cells via Innate Immunity. Our discovery that AID also functions outside the immune-system lead us to characterise the molecular aspects of Epigenetic Reprogramming and Meiotic Recombination. As an enzyme capable of inducing mutations and recombination it is self-evident that we also investigate how AID can induce Oncogenesis.
DNA deaminases are a family of proteins that deaminate cytosine bases in single stranded DNA (ssDNA). Originally thought to function as genome guardian to protect against foreign DNA (e.g. retroviruses, transposons etc.), the ancestral member, AID, has co-evolved with the immunoglobulin loci, and is essential for the formation of a functional humoral immune response . In humans, this family has significantly expanded and consists now of 8 members (also known as the APOBEC3 family) , most of which are part of the innate immune system. Our recent efforts, as well as those of other laboratories, have now identified a second function of AID: With its ability to deaminate 5-methyl-cytosine (5meC), AID may play a significant role in epigenetic reprogramming . Because the physiological function of DNA deaminases is to induce DNA damage, they are very powerful mutagens and may be involved in all stages of oncogenesis.
There are numerous pathways of how AID could become activated ?">, including transcription, translation, protein complex formation, etc. We have recently been able to show that AID can be directly activated by the hormone estrogen , while progesterone inhibits AID mRNA formation . This discovery has far reaching consequences, as it may have answered an over 100 year old question: How can Estrogen induce cancer? As estrogen itself is non-mutagenic, it could only act through an intermediate complex or activity. By directly activating AID transcription, estrogen is able to induce a DNA mutator in a variety of hormonally responsive cells.
Figure 2 . Schematic of estrogen induced activation of AID. Extracellular estrogen enters the cell, binds to the estrogen receptor dimer (ER), is transported as a ligand complex into the nucleus, binds near the AID promoter, and activates AID transcription.
This area of our research is one of the most important ones in the field of DNA deaminases. As mentioned, having a DNA mutator roam freely in the nucleus would be very 'dangerous' as it could mutate off-target loci and induce genome instability. Rather than using a hypothesis driven approach, we used classical biochemistry to understand which proteins are associated with AID on chromatin. This unbiased approach has allowed us to identify the RNA pol II elongation complex as the centre of a potential targeting mechanism. Although it may be counterintuitive, as RNA pol II is widely distributed throughout the genome, this discovery in combination with our work on AID's involvement in epigenetics (see below) and its direct link to DNA repair (also below), provides a framework for important research in the near future.
Although hydrolytic deamination is a relatively simple biochemical process, an oxygen replaces a NH2 group, the precise mechanism for DNA deamination is of significant relevance. We were able to show that the human DNA deaminases (AID and APOBEC3 family members) have a sequence preference surrounding the target dC (e.g. AID prefers AGC whereas APOBEC3G prefers CCC)   , which is a reflection of their physiological function. We could also show that AID acts in a distributive manner (e.g. between each deamination event AID dissociates briefly from the ssDNA target), and APOBEC3G in a processive manner (it stays associated with the target) . This difference in kinetics is important in terms of pathologies, as a processive nuclear AID is more oncogenic (multiple lesions in the same loci) than a distributive one. Currently we are investigating how the chemical reaction during catalysis (including binding to the substrate) proceeds, using chemically modified DNA bases.
AID was originally identified to play a key role in the acquired immune response. Our breakthrough came, when we discovered that AID directly mutated DNA - a clear break with the dogma that DNA has to be maintained in a stable configuration . Inducing dU lesions in the immunoglobulin locus initiates DNA repair pathways that do not result in repair, but mutation and recombination. Our work on hormonal activation of AID provides a possible explanation why B cell autoimmunities producing pathological antibodies, such as Lupus, are more predominant in women . Furthermore, aside from analysing the sequence specificity   , mode of action , or the involvement of the RNA pol II targeting complex, we have begun to dissect the molecular mechanisms of AID induced lesion resolution. Using an in vitro biochemical approach in cellular extracts (the first such system available), we have been able to show that post-lesion generation, AID recruits DNA repair proteins to the deaminated dC and alters repair functions - a novel activity of AID. We are currently analysing an AID interacting factor part of the DNA damage response pathway. The system also provides an insight into how single AID-induced events can lead to complex processive like alterations in the DNA.
Although our lab does not place a large focus on the involvement of DNA deaminases in innate immunity, we have been able to contribute to the field in the past and are currently still doing so. The biochemical and genetical analysis of the difference in sequence context   provided an insight of how APOBEC3G may target the HIV genome. The sequence preference of APOBEC3G is predominately found near the UTRs of HIV, a stretch of poly(C) needed for second strand synthesis. The processivity of APOBEC3G  is in a 3' to 5' direction, possibly due to the presence of a non-functional second deaminase domain . Currently, we have initiated experiments to determine if the effect of hormonal stimulation of DNA deaminase expression can also translate into an enhanced physiological efficacy.
If AID was limited to activated B cells, one could argue that DNA deamination was a phenomenon only necessary for a particular sub-set of cells (B cells), and its importance as a global protein would be limited. From the outset though, we considered DNA deamination to be part of a general genome guardian mechanism. Our discovery that AID is expressed in oocytes and other important developmentally regulated tissues, was hinting towards a second physiological function of AID. Combining this finding with our biochemical analysis showing that AID can deaminate 5meC in the context of a CpG , indicated that AID may play a role in epigenetic reprogramming. Our current work, in collaboration with Dr. Reik in Cambridge, UK, is centred around the precise molecular mechanisms that are needed to initiate AID induced de-methylation. Our in vitro system has again provided novel insight into this process and may provide a paradigm shift on how DNA demethylation is initiated - via AID's role in recruiting DNA repair proteins and inducing processive demethylation.
Although DNA deaminases directly deaminate DNA, cells also need to provide the right environment to process the lesions into DNA altering products. We therefore explored the effect of DNA deaminases in germ cells in organisms that do not naturally posses those enzymes. During meiosis DNA is broken and recombined to provide variation to offspring and induce chromosome segregation. By showing that AID was able to provide some of these functions to yeast and nematode , we demonstrated that DNA damage other than double strand breaks can induce meiotic recombination, and meiotic nuclei but not pre-meiotic nuclei process AID lesions into recombination substrates. This work also led us to establish a new set of powerful genetic tools that allow us to dissect the unimpeded targeting of AID in the absence of B cell specific factors.
Understanding the fundamental principles of any biological process is vital for science and society, yet at the same time one should never forget that more immediate concerns have to be addressed as well. DNA deaminases are the first proteins identified to have mutagenic activity as their physiological function. This changes the outlook of how to analyse any disease with genetic instability, as it is no longer the mutant protein that is causing the abnormality, but how the cell controls and regulates a physiological activity. Our work on hormones has shown that stimulating normal cells with physiological amounts of estrogen leads to oncogenic mutations and translocations . Moreover, our kinetic analysis showed that enzyme activity can be used to control potential oncogenic activity .
Figure 3. Estrogen induced c-Myc to IgH translocations. Estrogen induced hyperexpression of AID can induce off-target translocations, leading to DNA recombination between the IgH enhancer and the c-Myc promoter. This hyper transcription of c-Myc leads to oncogenesis.
 Petersen-Mahrt, S. (2005) Immunol Rev 203, 80-97.
 Conticello, et al. (2005) Molecular biology and evolution 22, 367-377.
 Morgan, H. D., et al. (2004) J Biol Chem 279, 52353-52360.
 Pauklin, S., et al. (2009) J Exp Med 206, 99-111.
 Pauklin, S., and Petersen-Mahrt, S. K. (2009) J Immunol 183, 1238-1244.
 Harris, R. S., et al. (2002) Mol Cell 10, 1247-1253.
 Beale, R. C., et al. (2004) J Mol Biol 337, 585-596.
 Coker, H. A., and Petersen-Mahrt, S. K. (2007) DNA repair 6, 235-243.
 Petersen-Mahrt, S. K., et al. (2002) Nature 418, 99-103.
 Pauklin, S., et al. (2009) Genetics 182, 41-54.
update: Feb 2013
Dr. Wolf Reik
Baraham Institute, Cambridge, UK
Dr. Bernardo Reina-San-Martin
IGBMC Illkirch, France
Prof. Primo Schär
University of Basel, Switzerland
Dr. Vincenzo Costanzo
LRI - Clare Hall, London, UK
(update: Oct 2010)