DNA N6-adenine methyltransferases in eukaryotes

The N6 methylated adenine base (N6A) in DNA is widely present in prokaryotes but relatively uncommon in eukaryotes. However, right from the earliest days, DNA N6A methylation  has been  identified in several eukaryotic lineages including ciliates, chlorophyte algae and dinoflagellates. In some species 0.5-10% of the adenines in the genome may be modified. Ciliates in particular have been studied as a model for the regulatory role of adenine methylation (Click here to read articles on the same).  Until recently, the identity of the eukaryotic adenine methyltransferases was a mystery. A study in plants, reported a potential adenine methyltransferase, but it turns out to be an RNA methylase of the TRM11 family.


The N6A methylases and the related N4C methylases can be distinguished by a characteristic sequence signature in the active site (loop after strand-4) of the form [NDS]PP[YFW] (see structure above). However, this signature is present in both RNA and DNA adenine methylases and several RNA methylases are highly conserved across life.  The classical DNA N6A methylases (N6A-MTase or Dam), appear to have emerged from the HemK-RsmC-RsmD clade of RNA methylases (see figure from our previous study). Most prokaryotic N6A DNA methylases are found in R–M systems, which have been widely disseminated via lateral transfer across distantly related lineages, but in some instances such as in the E. coli Dam protein and in the Caulobacter CcrM, they have been taken up for host function. N4C is another amine group base methylation seen in DNA, but sequence analysis reveals that they have emerged from the N6A methylases on multiple occasions independently.


In a recent study, we now report at least six distinct families of DNA N6-adenine methylases (N6A-MTase or Dam) in eukaryotes. Several of these are specified by different types of mobile elements. For example the Dictyostelium DIRS-1-like retrotransposon element encodes a N6A methylase. This element is widely disseminated across eukaryotes and expanded in several distantly-related organisms. However, except certain chlorophyte algae the methylases in most other genomes are inactive. This suggests that in most other species the methylase domain is likely to only function as a DNA-binding regulatory protein instead of a methylase. Another family found only in Trichomonas  includes N6A-DNA methylases that are often fused to phage structural proteins. The CrRem1-like LTR-containing retrotransposons of Chlamydomonas encode a polyprotein with the  methylase fused to C-terminal aspartyl protease and reverse transcriptase domains. Another predicted mobile element seen in chlorophytes and certain chythrid fungi contains a Dam methylase that is fused to a ParB-like HTH domain. The pervasive presence of Dam-like methylases associated with distinct groups of transposons suggests that they might act in cis to control their own gene expression and mobility through methylation of specific adenines within themselves or in their vicinity. 


We also report eukaryotic N6A methylases that appear to be cellular enzymes with a role in chromatin organization. One of these, found across chlorophyte algae is a multidomain protein with the N6A methylase domain fused to one or more N-terminal BMB/PWWP and C-terminal PHD-X/ZF-CW domains. Additionally, they often contain PHD finger domains N-terminal to the methylase domain. The accessory domains function as adaptors that read the histone code. The architectural syntax suggests that these enzymes localize to specific regions of chromatin where the histones bear modified marks such as trimethylated lysines, which is followed by localized N6A or N4C methylation.


Most of these eukaryotic methylases are obviously related to the DNA N6A-methylases of mobile elements of prokaryotes, suggesting their derivation from prokaryotic mobile elements. The situation in ciliates, which are a model system for studying the regulatory role of N6A methylation in eukaryotes is quite different and interesting. Sequence searches failed to reveal any obvious DNA N6-MTase candidates. However, ciliates  have a distinctive paralogous version of the Ime4-like family of adenine methylases, which is fused to N-terminal ZZ Zn-fingers, a domain of the treble-clef fold also found in chromatin proteins such as Ada2 and CBP/p300. Given that all ciliates studied to date show substantial N6mA in DNA, and have no other candidate methylases to catalyze this reaction, we suggest that these ZZ-domain containing methylases indeed perform this function. Orthologous methylases of this ciliate version are found in amoeboflagellates like Naegleria and the rhodophyte alga Cyanidioschyzon, suggesting a possibly wider distribution for this form of adenine methylation across eukaryotes. 


The Ime4/MunI-like methylases have several interesting aspects. They are circularly permuted. Bacterial versions are part of Restriction-modification systems, suggesting that these are DNA methylases. In contrast, the classical eukaryotic IME/MT-A70 methylases, which are derived from bacterial methylases, are RNA methylases. Although ciliates have the regular IME4-like version, the presence of this distinct paralog and its architecture suggests a return to DNA methylation in these species!!


You can read about this in our detailed study on DNA methylation-demethylation systems (Click here). 

Tête-à-tête with TETs-- Methylcytosine hydroxylation and demethylation

The report of the existence of a stable hydroxymethylcytosine modification in animals and the discovery of the enzymes that catalyze this modification represents one of the great triumphs of collaborative research. It has veritably opened a new angle to epigenetic regulation and our pioneering paper in the field has over 200 citations in just under two years. Scientific papers often conceal the history, drama and effort that goes into discovery and this study too had its fair share.

Our interest in this family was kindled by the discovery by Dr. Piet Borst's group of the enzymes that catalyze an unusual base modification in kinetoplastids. Euglenozoans and their parasitic relatives the kinetoplastids contain a distinct modified thymidine derivative in their DNA; base J. Unlike several phages, where modified nucleotides are incorporated during replication, base J conversion occurs on DNA and its synthesis was shown to involve two steps. First, thymidine is hydroxylated to hydroxymethyluracil and this is then glucosylated by a glucosyltransferase to give Base J.

Two proteins, JBP1 and JBP2, were shown to be important for base J synthesis and threading analysis revealed a common hydroxylase domain of the 2-oxoglutarate and iron dependent dioxygenases (2OGFeDO) superfamily in both of them, which suggested a 2OGFeDO-like mechanism that catalyzed the first step. You can read about the discovery of base J here.

Given our long-term interest in this superfamily (see an early study), we investigated this domain using sensitive sequence analysis methods and we noticed that it was also present in bacteria, phages, fungi, Naegleria and in the animal TET proteins (e.g. CxxC6), mutations in which were known to cause various myeloid cancers. Interestingly, in animals the domain was fused to the CXXC domain that is known to discriminate methyl cytosine containing nucleotides in DNA. This gave us a clue that the TET proteins modified DNA in situ like the JBP1/2 proteins, and perhaps at cytidine instead of thymidine. Given its hydroxylase function, we speculated that it might be involved in a DNA demethylation pathway through a hydroxyl intermediate, in a 2007 paper in the International Journal of Parasitology. However, there were too many unanswered questions and a hydroxylation-based C5-demethylation of cytosine was unprecedented in vivo (although it can be achieved in vitro), as compared to N-6 demethylation through a hydroxylase intermediate (e.g. as in the AlkB reaction). Experimental investigations, thus, were key to further understanding the biochemistry and functional role of this domain in animals.

This was when Dr. Anjana Rao's group at Harvard joined the effort in trying to unravel the function of the animal homologs containing the JBP1/2-like dioxygenase domain. Dealing with these proteins is no joke as they are gigantic. They range in size from 1600-2300 amino acids. Further, TET1, TET2 and TET3 contain a distinct cysteine rich domain inserted into the N-terminal part of the 2OGFeDO domain and TET1 and TET3 also contain an N-terminal CXXC domain which is separated from the cysteine-rich and 2OGFeDO domains by a long low complexity region  (Click here for a text alignment or here for a colored version). The heroic efforts of Mamta Tahiliani resulted in the purification of these proteins and the development of an in vitro system to assay their activity. Overexpressing TET1 or just a minimal version with the 2OGFeDO domain showed a decrease in 5mC levels and the presence of an distinct TLC band. This was also reproducible in the in vitro assay. In collaboration with Dr. David Liu's group the distinct nucleotide species was shown to be hydroxymethylcytosine. Hydroxymethylcytosine research was prominent in the heydays of phage biology, it now had a second birth.

The study was published in two parts that were released almost simultaneously. A comprehensive theoretical analysis of the superfamily, its phyletic distribution, evolutionary history and contextual associations was published in Cell Cycle and the experimental studies were published in Science. TET proteins are now implicated in embryonic stem cell re-programming, cell lineage specification, reprogramming the paternal genome, fine-tuning transcription and opposing aberrant methylation. Our analysis provided the first biochemical leads towards understanding the basis of myelomas caused by mutations in the TET proteins. The list is only going to increase over time. While the animal proteins have received much press, there are several gems in the waiting in other eukaryotes. More on this in the next post.