Methyltransferase‐Directed Labeling of Biomolecules and its Applications

Abstract Methyltransferases (MTases) form a large family of enzymes that methylate a diverse set of targets, ranging from the three major biopolymers to small molecules. Most of these MTases use the cofactor S‐adenosyl‐l‐Methionine (AdoMet) as a methyl source. In recent years, there have been significant efforts toward the development of AdoMet analogues with the aim of transferring moieties other than simple methyl groups. Two major classes of AdoMet analogues currently exist: doubly‐activated molecules and aziridine based molecules, each of which employs a different approach to achieve transalkylation rather than transmethylation. In this review, we discuss the various strategies for labelling and functionalizing biomolecules using AdoMet‐dependent MTases and AdoMet analogues. We cover the synthetic routes to AdoMet analogues, their stability in biological environments and their application in transalkylation reactions. Finally, some perspectives are presented for the potential use of AdoMet analogues in biology research, (epi)genetics and nanotechnology.


Introduction
Biological systems are complex and never merely the sum of their basic components.I nstead, the intricate interplay between countless numbers of biomolecules gives rise to new, "emergent" properties that could not have been predicted when considering each component individually. [1] In this respect, life itself is perhaps the most intriguing emergent property of all. Nonetheless,e ven at iny change in even asingle macromolecule can have profound effects on aliving organism. Such is the case in the epigenetic regulation of gene activity,w here small, highly specific modifications to DNA, RNA, or interacting proteins will result in significant modification of biological function.
Chemists can utilize the inherent specificity of the enzymatic machinery involved in epigenetic regulation to deliver functional or reporter groups to defined macromolecular targets.
This review will focus on as ubset of methyltransferase (MTase) enzymes,t he S-adenosyl-l-methionine (AdoMet 1, Scheme 1) dependent methyltransferases.I nn ature,t hese enzymes methylate their targets,w hich range from small molecules to proteins and oligonucleotides.They perform this function by binding the methyl donor AdoMet (1)a nd catalyzing the covalent transfer of am ethyl group from AdoMet to their target. Moreover,m any MTases have been shown to be sufficiently malleable that they will perform as imilar catalytic transfer of much more extended and functional moieties than the methyl group.T his has allowed them to become an essential tool in the functionalization of the three major biopolymers:RNA,DNA,and proteins.T he field has been the focus of increasing attention and over the past decade,s everal new approaches that allow the incorporation of functional chemical moieties into biomolecules in asite-specific manner have been developed.

Labeling Strategies Using AdoMet-Dependent MTases
AdoMet-dependent MTases catalyze the transfer of am ethyl group from their ubiquitous cofactor to at remendously diverse range of biomolecules (Scheme 1). This makes Methyltransferases (MTases) form alarge family of enzymes that methylate adiverse set of targets,r anging from the three major biopolymers to small molecules.Most of these MTases use the cofactor Sadenosyl-l-Methionine (AdoMet) as amethyl source.Inrecent years, there have been significant efforts towardthe development of AdoMet analogues with the aim of transferring moieties other than simple methyl groups.T wo major classes of AdoMet analogues currently exist:doubly-activated molecules and aziridine based molecules,each of whichemploys adifferent approacht oa chieve transalkylation rather than transmethylation. In this review,w ed iscuss the various strategies for labelling and functionalizing biomolecules using AdoMet-dependent MTases and AdoMet analogues.W ecover the synthetic routes to AdoMet analogues,t heir stability in biological environments and their application in transalkylation reactions. Finally,s ome perspectives are presented for the potential use of AdoMet analogues in biology research, (epi)genetics and nanotechnology. them fundamental to aw ide variety of biological pathways, including small-molecule biosynthesis,p rotein repair, signal transduction, chromatin regulation, and gene silencing. [2] In order to catalyze methylation, the MTases form ternary complexes with their target molecule and AdoMet. These complexes involve an etwork of intermolecular interactions, but some general principles are common to all systems.T he transferable methyl group of AdoMet is bound to asulfonium center. Them olecule is inherently unstable towards nucleophilic attack and the methyl group readily participates in substitution reactions. [3] This results in ah alf-life of 17 hours for AdoMet in the M.HhaIreaction buffer at pH 7.4, 37 8 8C, [4] and this susceptibility to nucleophilic attack is instrumental to the function of MTases.Amechanism for C5 cytosine methylation by the HhaIm ethyltransferase (M.HhaI), which results in 5-methyl cytosine (5mC), was first described by Santi et al. [5] Thea uthors proposed at wo-step concerted mechanism (Scheme 2), which is an enzymatically catalyzed version of the Morita-Baylis-Hillman reaction. [5a] In the first step of the reaction, cytosine undergoes nucleophilic attack at C6 by the thiol of acysteine residue in the catalytic pocket of M.HhaI. This results in the 6-Cys-cytosine compound (C5A). Irreversible nucleophilic attack at the transferable methyl group by C5A results in the stable intermediate 5-methyl-6-Cys-5,6-dihydrocytosine (MCD) with simultaneous conversion of AdoMet into AdoHcy. [5b] Deprotonation and subsequent release of the enzyme of MCD restores the aromatic conjugation and results in the DNA5 -methylcytosine (5mC) product. [5a] In N6-adenine DNAM Tases,t he 6-amino group of adenine forms two hydrogen bonds,w hich increases the electron density of N6 and contributes to its activation for nucleophilic attack.
This mechanism is not specific to M.HhaIa nd has been shown to apply to many other MTases. [3,6] An S N 2mechanism is proposed at the sp 3 carbon center,w ith S-adenosyl homocysteine (SAH) acting as ag ood and stable leaving group.T his has proven instrumental in the development of alarge variety of artificial AdoMet analogues.

AdoMet Analogues
AdoMet is as tructural hybrid of methionine and adenosine for which both the R and the S diastereoisomers bind similarly to MTases.H owever,o nly the S diastereoisomer has the correct geometry at the sulfonium center to allow proper catalytic function. [7] In nature,the active S diastereoisomer is formed through the stereospecific reaction of lmethionine (l-Met) with adenosine triphosphate (ATP), which is catalyzed by the methionine adenosyltransferase (MAT) enzymes. [8] Thev ast majority of artificial AdoMet analogues can be categorized into two major groups:a ziridinoadenosines and doubly-activated AdoMet analogues (Scheme 3). To date,the aziridinoadenosines have only been prepared synthetically. However,t he doubly-activated AdoMet analogues can be prepared either through total synthesis or enzymatically, using asynthetic methionine analogue and aMAT enzyme. [9] 2.1.1.

Aziridine-Based AdoMet Analogues
Aziridine-based AdoMet analogues (Scheme 4) constitute some of the earliest examples of AdoMet analogues and were first developed by the Weinhold group.
[10] These AdoMet analogues differ from natural AdoMet in that their 5'-sulfonium is replaced by an aziridine ring (11,S cheme 8). This aziridine group undergoes ring opening as ar esult of nucleophilic attack by the target compound. Since the "leaving group" is conjugated to the transferable moiety in this case,following the S N 2reaction mechanism to completion sees the transfer of the entire cofactor analogue to the target molecule (DNAi nS cheme 3). [7] Several DNAM Tases were found to be capable of transferring these compounds to DNA. Thep rocess of introducing reporter groups (e.g.,b iotin (3),S cheme 4) to av ariety of DNAs equences has been termed "sequencespecific methyltransferase-induced labeling" (SMILing). Many different reporters,f or example,f luorescent dyes have been attached to the adenine moiety of this aziridine AdoMet analogue for application in DNAl abeling. [10b, 11] Modelling of the AdoMet binding pocket from crystallographic data showed that steric interactions between the cofactor analogues and the enzyme are likely ak ey factor determining compatibility of AdoMet analogues with specific methyltransferases.A sa ne xample,m odelling of the TaqI methyltransferase (M.TaqI) AdoMet binding pocket showed that the 8-position of the adenine moiety of abound cofactor is accessible to the solvent. This implies that the enzyme is therefore likely to tolerate cofactor analogues that incorporate bulky modifications at this site.S uccessful examples of such bulky modifications to the cofactor analogue include an azide group (12,T able S1 in the Supporting Information) and ar ange of fluorophores. [11][12]  similar opportunities for modification of the adenine moiety (6-, 7-or 8-position) for ar ange of MTases from their crystallographic structures. [10] Further work on aziridine-type AdoMet analogues has focused on the preparation of 5-N-substituted nitrogen mustard (N-mustard) compounds rather than direct synthesis of the corresponding aziridine.These N-mustards were found to be reasonably stable and have the ability to form an aziridinium ion in situ (Scheme 4). This approach largely avoids the synthetic difficulties associated with the inherently unstable aziridine moiety. [13,14] In addition, 5-N amino acid substituted N-mustards (13,T able S1) were shown to be significantly more active in MTase directed transalkylation of DNAc ompared to their counterparts featuring alkyne substituents.S imilarly,i nclusion of af unctional group such as an azide (14, 15 in Table S1) or at erminal alkyne (4 in Scheme 4, 16 and 17 in Table S1) enables the use of these Nmustard cofactor analogues in bioorthogonal ligations. [12,13] Theu se of the aziridine-based cofactor analogues for transalkylation reactions suffers from several drawbacks. Most notably,t he transalkylation is not catalytic,a nd stoichiometric amounts of the MTase must be used in the reaction. This was first observed by Osborne et al.,w ho demonstrated the protein methyltransferase 1( PRMT1)catalyzed transalkylation of ap eptide using an N-mustardbased cofactor. Thep roduct of this reaction, essentially ap eptide with ac ovalently bound cofactor analogue,i s apotent inhibitor for the MTase. [15] Hence,the reaction is selflimiting. Furthermore,the aziridine-based analogues are very reactive.T his makes them prone to rather rapid degradation and also results in an increased propensity to effect nonspecific alkylation, even in the absence of MTases. [7] 2.1.1. Doubly Activated AdoMet Analogues MTases have an inherent ability to catalyze the transfer of alkyl groups larger than methyl groups.H owever,f or DNA MTases,t he transfer rate decreases rapidly with increasing size of the transferable moiety (methyl > ethyl > propyl). The reaction occurs through an S N 2m echanism with inversion of the configuration at the a-carbon next to the sulfur atom. [16] As such, incorporation of an unsaturated bond at the b-position can potentially stabilize the p-orbital of the a-carbon formed at the intermediate stage of this substitution reaction. Weinhold, Klimašauskas, and co-workers were able to demonstrate that an allylic or ap ropargylic carbon-carbon bond at the bposition relative to the sulfonium center can restore the reaction rate through conjugative stabilization of the S N 2t ransition state (Scheme 5). [17] This approach to labeling,using so-called "doubly-activated" cofactor analogues,h as been termed "methyltransferase-directed transfer of activated groups" (mTAG). [18] Unlike aziridine-based labeling,m TAGi s at ruly catalytic process and does not require stoichiometric amounts of the MTase. [17b] Even though multiple turnovers per minute can be achieved, the process is still typically an order of magnitude slower than the equivalent methyl transfer. [17b] AdoMet analogues with alkyl, alkenyl (19,T able S2), and alkynyl (20,T able S2) side chains have been prepared and successfully applied in mTAG. [17] Thea pproach offers as imple way to label targets with af unctional group that can be used for further ligation of more complex moieties.F or example,incorporation of different transferrable amine moieties (21, [4,18] 22, [4,19] 23, [4,19b] Table S2) can be followed by ac oupling reaction with the N-hydroxysuccinimide (NHS) ester of an amine-reactive probe. [18] In addition, several examples exist on the application of AdoMet analogues with at ransferrable terminal alkyne (5, [20] Scheme 6; 20, [20g,h] 24, [20b,c,g] and 25, [4,19b] Table S2) or azide (26 g, [20,21] 27). [4] Thet ransferred alkynes and azides can be further functionalized by using the biocompatible and highly-efficient azide-alkyne cycloaddition reaction (one of the "click" series of reactions). However,s ome of these analogues,s uch as compound 20 (Table S2), which features at erminal alkyne,a re highly unstable. [4, 20g,h] Other AdoMet analogues,s uch as the AdoEnyYn compound (5,S cheme 6), are more stable,t hus making them more suitable for use in bioorthogonal ligations. [20b-h] Ketones are yet another interesting transferrable moiety since they can be used to react with hydroxylamines and hydrazides.This was demonstrated with AdoMet analogue 28 (Table S2), which was successfully transferred to DNAa nd subsequently used for coupling with ahydroxylamine coupled to af luorophore. [22] Finally,A doMet analogues directly incorporating af luorescent dye allow single-step direct labeling reactions.A ne xample in which aT AMRA dye was coupled to the AdoMet analogue (6,S cheme 6) and subsequently used for labeling reactions was first described by Grunwald et al. [23]

Stability of AdoMet Analogues
AdoMet analogues are prone to spontaneous decomposition in aqueous environments through ar ange of different pathways (Scheme 7). [4,24] Reversible racemization (route a) to the R diastereomer of AdoMet is also common and can result in the formation of an inactive AdoMet analogue.T he presence of the sulfonium center activates the adjacent carbon atoms towards decomposition reactions.U nder alkaline conditions,d eprotonation at the 5'-C occurs (route b), which results in formation of adenine and S-ribosylmethionine.Under more acidic conditions,intramolecular attack by the a-carboxylate group on the g-C of the methionine group (route c) yields methylthioadenosine (MTA) and homoserine lactone (HSL). In the case of AdoMet, no nucleophilic attack is observed at the methyl group,b ut this changes when considering AdoMet analogues with extended chains containing at riple bond in the b-position relative to the sulfonium center.T he partial positive charge at the 1''-C increases,thereby making it more susceptible to nucleophilic attack (route d). When electronegative groups (e.g.,Y = NH 2 )a re in close proximity to the triple bond, this can lead to ah igher electron deficiencya tb oth 4''-C and 1''-C,w hich enables the addition of water (route e) to both to give the hydrated compound, which shows almost no reactivity towards most DNAMTases and protein MTases. [4,24] At pH 7.5 and 37 8 8C, the rate constants for racemization, cleavage to HSL/ MTA, and hydrolysis to adenine/S-ribosylmethionine were reported to be 1.8 10 À6 s À1 ,4 .6 10 À6 s À1 , and 3 10 À6 s À1 ,r espectively.T he hydrolysis rate shows asignificant decrease at lower pH values. [25] Recently,H uber et al. developed novel AdoMet analogues lacking the N7 of the adenine moiety and further omitting the carboxylic acid of the amino acid chain in favor of atetrazole ring. This makes them less likely to decompose through pathways band cand restricts epimerization at the sulfonium center. [26] To improve the stability and reactivity of the AdoMet analogues,s everal molecules have been studied in which sulfur is replaced with another chalcogen, such as selenium (SeAdoMet (7), Figure 1) or tellurium (TeAdoMet (8), Figure 1). ForA doMet, three decomposition pathways are common:r acemization to the inactive diastereomer,d eprotonation of the 5'-C to yield S-ribosylmethionine,and adenine or intramolecular nucleophilic attack by the carboxylate group to give HSL and MTA. SeAdoMet (7,F igure 1) was found to decompose via two of these pathways,w hereas the tellurium analogue Te AM (8,F igure 1) was inert to all three pathways. [24] It has been shown that the trend in electrophilicity of these AdoMet analogues follows the order SeAdoMet > AdoMet > Te AdoMet and that the 5'-C acidity follows the order AdoMet > SeAdoMet > TeAdoMet. [27] Selenium analogues are therefore both more reactive towards nucleophilic attack, thus making them better suited as cofactors for transalkylation, and more stable towards deprotonation of the 5'-H (pathway b, Scheme 7). As ar esult, they have been subject of extensive investigation. Weinhold et al. have developed SeAdoYn (9,F igure 1) as ac ofactor for protein-methyltransferase-mediated fluorescent labeling of proteins.T his analogue showed improved

Angewandte Chemie
Reviews reactivity and higher stability toward hydrolysis compared to the similar sulfur-containing AdoEnYn and AdoYn analogues. [20h] Comparative decay studies with the propargylic SeAdoYn (9, Figure 1) and ProSAM (20,T able S2) analogues showed that decomposition of the ProSAM compound first results in the hydrated product (keto byproduct 28,T able S2), which further decomposes to the thioether.H owever,S eA-doYn follows ad ifferent mechanism and directly forms the selenoether. [20h, 28] SeAdoYn is used as as ubstrate for transalkylation by alarge variety of wild-type MTases,which is in contrast with many of the AdoMet analogues with larger transferable groups,w hich are only active with mutated MTases. [9b,19a, 20h, 28, 29] TheL uo group has also demonstrated that selenium substitution can result in increased reactivity compared to the sulfur-centered AdoMet analogues when applied as sub-Scheme 7. Decomposition pathwayso f(S,S)-AdoMet (and its analogues). a) Inversion at the sulfoniumcenter of (S,S)-AdoMetresults in (R,S)-AdoMet. b) Deprotonation at the C-5' and subsequent elimination of adenine base to give S-ribosylmethionine. c) Nucleophilic attack by the acarboxylate on the g-carbon of methionine delivers HSL and MTA. d) Nucleophilic addition at 1''-C. e) Addition of water to the 2''-C or 4''-C position.

Angewandte Chemie
Reviews 5188 www.angewandte.org strates for the protein MTases.T heir study shows that the decomposition rate of SeAdoMet is 10-fold higher than that of AdoMet, but the protein-MTase-catalyzed transalkylation reaction is only 3-5-fold faster for the SeAdoMet analogue. Hence,the authors suggest that the protein-MTase-catalyzed reaction rates are not determined solely by the strength of the chalcogen-carbon bond, but are likely caused by other factors. [28,30] Interestingly,t he study also suggests that the bsp 2 carbon atom, which is essential for activity with S-alkyl AdoMet analogues,i sl ikely not required for some protein MTases when Se-alkyl AdoMet analogues are used. [30] 2.3.

Synthesis of AdoMet Analogues
Thef irst example of the synthesis of aziridine-based AdoMet analogues was described in 1998 by Pignot et al., who focused on the development of aziridine-based cofactor analogues. [10a, 31] N-Adenosylaziridine was synthesized through nucleophilic substitution of the tosylate group in 5'deoxy-5'-tosyladenosine with aziridine and subsequently activated as an alkylating agent through protonation of the nitrogen atom in the aziridine ring (Scheme 8). [10a] Thering strain in this three-membered heterocycle makes it susceptible to nucleophilic ring opening, which is further facilitated by nitrogen quaternization. This aziridine-based AdoMet analogue can be used for the introduction of different functional or reporter groups through modification of the adenine moiety.O ne such example shows the attachment of adansyl fluorophore at the adenine 8-position. [11] The synthesis starts from 8-bromo-2',3'-O-isoporpylideneadenosine,w hich was treated with diaminobutane to introduce aflexible linker and was converted in afew steps (substitution and deprotection) into the N-adenosyl aziridine derivative.A similar route has been used to prepare the N-mustard precursors of the aziridine-based AdoMet analogues (Scheme 4). [32] These aziridine-based AdoMet analogues suffer from al ong and low-yielding synthetic route. [10,12b,13, 14] Townsend et al. have reported improved synthetic routes for the synthesis of N-chloromustard-substituteda denosines using reductive amination as the key step and have also demonstrated the synthesis of ap hotocaged derivative,w hich benefits from increased stability compared with standard aziridine-based cofactors and is readily activated as acofactor through UV irradiation. [33] Doubly-activated AdoMet analogues are typically synthesized by combining S-adenosyl-l-homocysteine (AdoHcy (2), Scheme 9) with an excess of strong electrophiles,f or example,a lkyltriflates or alkylbromides,u nder acidic conditions. [34] Thefirst chemical synthesis of the diastereomers of AdoMet from AdoHcywas published in 1959. [35] In an acidic environment, the propensity for nucleophilic attack by AdoHcy amines,h ydroxy groups,o rt he carboxyl acid, for example,isstrongly reduced, leaving the thioether as the sole reactant. [17a] Both the R and the S epimers are formed during this S N 2reaction. However,since only the active S epimer can be used in transmethylation reactions,aseparation of the diastereomers is preferable.R eversed-phase HPLC (RP-HPLC) is commonly used in efforts to separate the diastereomers.T his process,however, remains challenging.
Thesynthetic routes to AdoMet analogues are not trivial. In addition, these compounds often display limited stability. As aresult, there have been several attempts have to prepare both AdoMet and its analogues enzymatically.
AdoMet itself can be obtained through isolation from yeast grown in media supplemented with l-methionine. [8b] Small-scale [8a, 36] (mmol) enzymatic syntheses of AdoMet from ATPa nd l-methionine (l-Met) as well as larger scale [8b] (mmol) syntheses have been reported. TheA doMet formation is catalyzed by l-methionine S-adenosyl transferase (MATorAdoMet synthetase) in atwo-step fashion where the complete tripolyphosphate (TPP) chain of ATPiscleaved and subsequently hydrolyzed to pyrophosphate (PPi)a nd phosphate (Pi). Thee nzymatic synthesis results in high yields of the preferred epimer (À)-AdoMet (Scheme 10, top). [36a, 37] Recently,s everal groups have demonstrated enzymatic syntheses of AdoMet analogues as well. [9,38] TheT horson group tested ar ange MATe nzymes for their ability to catalyze the formation of several AdoMet analogues.T hey tested al ibrary of 44 non-native S/Se l-Met analogues with five different MATs.O ft hese,h uman MATIIw as the most permissive.Furthermore,the synthesized analogues could be successfully used in the alkylation of small molecules (indolocarbazoles). [9b] TheBurkart group have demonstrated the chemoenzymatic synthesis of several AdoMet analogues using either af luorinase (FDAS) from Streptomyces cattleya or ac hlorinase (SalL) from marine bacterium Salinaspora tropica. [38,39] Both halogenases are known to catalyze the breakdown of AdoMet to give l-Met and fluoro-5'-deoxyadenosine (FDA) or 5'-chloro-5'-deoxyadenosine (ClDA). However,t he reaction can be reversed at low chloride/ Scheme 8. Synthesis of the N-Adenosylaziridine AdoMet analogue.

Angewandte Chemie
Reviews fluoride and high l-Met concentrations (Scheme 10, bottom). This enzymatic pathway can also be exploited by using various l-Met derivatives for the enzymatic synthesis of AdoMet analogues. [9a,38, 40] However, it should be noted that decreased activity of the l-Met analogues with increasing size was observed. [40] 2. 4

. Methyltransferases
Methyltransferases are typically categorized into five distinct families (classes I-V) based on structural similarities. [2a] Of these,Class IMTases are by far the largest group. They catalyze the majority of methylation reactions and include all DNAMTases and some protein MTases and RNA MTases. [2a] Thel argest group of protein MTases are the protein lysine MTases,and together with the protein arginine MTases,t hey play an important role in histone modification and thus ultimately in gene transcription. Most of the work in the past decade using modified AdoMet analogues has focused on class IM Tases and class VM Tases,w hich are also the focus of this Review.O ther,l ess common, classes include the MetH reactivation domain MTases (class II), precorrin-4 MTases (class III), and the SPOUT family of RNAM Tases (class IV). [2] While the MTases share little sequence similarity,they do share ahighly conserved structural fold. Thecore element of this conserved fold consists of seven-stranded b-sheets with three helices on each side,astructural feature commonly referred to as the Rossman fold. [2b,41] This core is shared among MTases that act on all different substrates,r anging from small molecules to DNAa nd proteins. [3,42] This structural similarity suggests that the labeling reactions with AdoMet analogues are likely universal and applicable across the entire range of MTases.
In labeling reactions,w ildtype MTases are most frequently used in combination with an AdoMet analogue. However,insome cases,this is not possible because the native fold of the MTases is incompatible with the AdoMet analogue.T his could be due to bulky groups on the AdoMet analogues that prevent agood steric fit or disruption of the needed binding interaction to the AdoMet binding pocket. In these scenarios,i tc ould prove useful to engineer the AdoMet binding pocket for amore favorable interaction with the AdoMet analogues,f or example by directed evolution of the MTase. [43] In other cases,itmight be preferable to use an MTase that interacts more readily with the AdoMet analogue rather than the natural AdoMet, for example,w hen using the labeling reaction in live cells in the presence of natural AdoMet. In this scenario,i tw ould be useful to engineer the AdoMet binding pocket such that the enzyme prefers the AdoMet analogue over the natural AdoMet. To achieve this,o ne possibility would be to apply a" bump-and-hole" strategy, where point mutations are introduced at the protein active site.S ubsequent screening of these mutants can then reveal those variants which display enhanced selectivity towards the AdoMet analogue rather than Adomet. [44] 2.4.1.

DNA Transalkylation
In bacteria, DNAM Tases play ar ole in defense of the bacterium against viral invasion by enabling ad istinction to be made between the host genome and invading viral DNA. [6,45] In eukaryotic cells,t he main role of DNAM Tases is the regulation of genes. [6] DNAm ethyltransferases can be subdivided depending on their target for modification (cytosine C5, cytosine N4, or adenine N6). Of these groups,t he cytosine C5 and adenine N6 MTases are found in many species of fungi, bacteria, and protozoa, while the cytosine N4 MTases occur only in bacteria. [6,46] TheD NA MTases typically recognize short, palindromic DNAsequences (2-8 bases long). Ordinarily,the target base for methylation lies buried within the DNAh elix and as ar esult, the DNAM Tases extrude this base from the DNA duplex, flipping it into the enzyme catalytic pocket, where transmetylation occurs (Figure 2). [47] Theb yproduct of this catalysis, S-adenosyl-l-homocysteine (AdoHcy( 2), Scheme 9), is subsequently released from the AdoMet binding pocket and, in cells,d igested by an AdoHcy hydrolase enzyme. [48] DNAM Tases were the first MTases that were shown to catalyze transalkylation reactions using AdoMet analogues. Part of the reason for this is likely the ease with which the progress of the labeling reaction can be followed. Every bacterial MTase has ap artner restriction endonuclease that Scheme 10. Enzymatic AdoMet synthesis using SalL and MAT.

Angewandte Chemie
Reviews 5190 www.angewandte.org cleaves DNAa tt he same site targeted by the MTase. However,i ft hat site is methylated (or alkylated), cleavage is prevented. This can be exploited to readily confirm the activity of the methyltransferase with the AdoMet analogue using gel electrophoresis.
Thef irst example of at ransalkyation reaction using an AdoMet analogue and aD NA MTase was the aziridine transfer with the M.TaqID NA MTase in 1998. [10a] Many aziridine analogues have subsequently been successfully transferred to DNA, for example,b yt he adenine DNA MTases M.EcoRI [12b] and M.BsecI [49] and cytosine DNA MTases such as M.HhaI [12a] and M.SssI. [14a] In these cases, functional moieties (including fluorophores) [11] were placed at various locations of the aziridine cofactor. Thea forementioned study by Pljevaljčić  As mentioned in Section 2.1.1, the aziridine AdoMet analogues suffer from some disadvantages,which has inspired Weinhold, Klimašauskas,a nd co-workers to develop the doubly-activated AdoMet analogues.T hese cofactors have been employed in successful transalkylation reactions with several DNAM Tases,w ith either adenine or cytosine as the target base.Acomprehensive overview can be found in Table 1. Thea ttachment of afunctional or fluorescent group to the DNAcan be achieved using asecond chemical coupling reaction, such as the coupling of amines to NHS esters [19c] or click-chemistry-based conjugation. [20f] Alternatively,l ike the aziridine cofactors,the fluorescent group can be synthetically coupled to the cofactor,thereby allowing one-step transfer of the fluorescent groups to the DNAs ubstrate.T his approach was first demonstrated with the DNAM Tase M.TaqI. [23] Ther ates transalkyation reactions with the doublyactivated AdoMet analogues have been dramatically improved by engineering of the cofactor binding pocket of the cytosine C5 MTases. [19d] Three amino acid side chains were selected based on their potential steric interaction with the cofactor, and replaced with shorter residues (two residues were replaced with alanine,w hile one residue was replaced with the smaller polar residue serine). Them utants showed as ignificant improvement in both binding affinity and transfer rate with the AdoMet analogues,r elative to the wild-type M.HhaI. In general, this effect was more marked for the longer cofactor analogues with transferrable groups with longer alkyl-chains than those with shorter chains.F urthermore,t he mutant enzymes show as ignificant reduction in methylation rate with the natural cofactor AdoMet. The binding constant of the enzyme with the natural cofactors AdoMet and AdoHcy was reduced as well. Thew eaker binding of AdoHcy may contribute to the increased catalytic efficiency of the AdoMet analogues.
Tw oo ft hree mutations were located within the highly conserved sequence motifs of M.HhaIa nd were readily mapped to locations on other C5 DNAM Tases.I ndeed, the same mutations were successfully applied to the cytosine-5 MTases M.HpaII and M2.Eco31I, and later to the CpG-

.2. RNA Transalkylation
In this section, we present ab rief overview of the application of the RNAm ethyltransferase enzymes that have been shown to catalyze transalkylation of RNAu sing AdoMet analogues.E arly studies have shown some promise in this area but all have been realized in vitro using synthetically prepared or in vitro transcribed RNAs ubstrates. Whether these methyltransferases can replace existing antibody-based approaches for the targeted labelling or capture of RNA-based substrates remains an area for further investigation.
Thefirst example of an RNA-methyltransferase-mediated transalkylation was with the tRNAm ethyltransferase Tr m1. [20d] This enzyme has been shown to catalyze the transalkylation of the N2 of guanosine 26 in tRNA Phe using ad oubly-activated SAM analogue with an alkyne functionality.T he modified tRNAw as subsequently fluorescently labelled using copper-catalyzed azide-alkyne cycloaddition (CuAAC; one of the "click" reactions) in order to generate fluorescently labeled tRNA.
Using in vitro reconstitution of an archaeal box C/D small ribonucleoprotein RNA2 '-O-methyltransferase (C/D RNP) Tomkuvienė et al. were able to demonstrate transalkylation of in vitro transcripts of tRNAa nd pre-mRNAm olecules. TheC/D RNP complex includes aguide RNAmolecule that is used to direct the specificity of the C/D RNP to these nonnatural substrates.A gain, in as econd reaction, the alkylated RNAm olecules were fluorescently labeled by CuAAC. [19a] Functionalization of both miRNAa nd small interfering RNA( siRNA) has been described by Plotnikova et al. [19b] Here,t he HEN1 2'-O-methyltransferase from Arabidopsis thaliana was used to direct transalkylation to the 3'-terminal nucleotides of small double-stranded RNAmolecules.Infact, this study demonstrates the direct transfer of biotin to the 3' ends of small RNAduplexes and their subsequent isolation using streptavidin-coated beads. [19b]

Protein Transalkylation
Thet wo most common types of AdoMet-dependent protein MTases are protein arginine MTases and protein lysine MTases. [53] However,t here are some protein MTases that target other sites such as other peptidyl chains or the No rC termini of proteins. [54] Them ain targets of arginine and lysine protein MTases are histones,f or which methylation of the histone is associated with repression of transcription. [55] Besides histones,n umerous other proteins have also been targets of methylation, where they play arole in many physiological pathways such as signal transduction and protein translocation. [56] Because of the role of histone methylation in the repression of transcription, protein MTases have been implicated in cancer, neurodegenerative diseases, and other diseases. [57] One early study in 2001 demonstrated that amutant of the yeast MTase Rmt1was selectively inhibited by N 6 -substituted AdoMet analogues.T his was the first successful demonstration of the bump-hole technique as am eans to develop selective combinations of mutant methyltransferase enzymes and tailored AdoMet analogues. [44] Thea ziridine-based cofactor analogues and the PRMT1 protein methyltransferase have been successfully employed for protein transalkylation. [15,58] Thed oubly-activated (mTAG)c ofactors were first employed in protein transalkylation reactions by Peters et al. [20e] This and al ater study describe the transfer of an alkyne group to the target of the histone H3K9 protein MTases Dim-5 and SETDB1. [59] TheL uo group has developed combinations of several AdoMet analogues and engineered protein MTases to enable bioorthogonal targeting of transalkylation reactions in cells. Their work is covered in detail in ar ecently published review [7] but we provide ab rief overview here.T he focus of their work has been the human protein MTases G9a (also known as EuHMT2), GLP1 (also known as EuHMT1), and PRMT1 (for ac omplete overview,s ee Table 2). Enzymemediated transalkylation reactions using an azido-modified AdoMet analogue with engineered G9a and GLP1 [21,60] and an alkyne-modified AdoMet analogue with engineered G9a [20c] and PRMT1 have been demonstrated. [20g] Additionally,aselenium-based SAM analogue with an alkyne linker was effectively employed in transalkyation reactions directed and catalyzed by the native protein MTases GLP1, G9a and SUV39H2. [28,60,61] An extensive study with eight native enzymes showed limited transalkylation by the three wild-type protein MTases G9a, GLP1, and SUV39H2 when using an allyl AdoMet analogue,and acomplete absence of activity with the bulkier AdoMet analogues. [62] However, as ingle mutation in these three enzymes was shown to have adramatic impact on their

Angewandte Chemie
Reviews 5192 www.angewandte.org ability to catalyze transalkylation reactions and, furthermore, led to as ignificant reduction in the rate at which they performed methylations using AdoMet. These mutants were thus found to be capable of performing transalkylation reactions with synthetic AdoMet analogues in the presence of native AdoMet (e.g., in cell lysates). [63] In as ubsequent study,I slam et al. investigated the function of the mutations Y1211A in EuHMT1 and Y1154A in EuHMT2, which further improved the enzymatic transalkylation rates when using the AdoMet analogues.T he residues targeted for mutation were identified as gatekeeper amino acids that had blocked access to the cofactor binding pocket for the bulkier, synthetically prepared AdoMet analogues. [20b]

Small-Molecule Transalkylation
Small-molecule or natural product methylation is ubiquitous across all branches of the tree of life.M any AdoMetdependent MTases that target natural products (NP MTases) exist and they can generally be categorized based on their preference for oxygen, nitrogen, sulfur,o rc arbon atoms as their nucleophilic substrates. [65] As such, NP MTases participate in the modification of al arge numbers of structurally diverse small organic molecules,w hich affects their bioavailability,a ctivity,a nd reactivity in processes ranging from metabolism to signaling and biosynthesis. [65b] Many NP MTases have been shown to feature aR ossman-like fold structural motif that is often extended by additional domains. These additions to the core enzyme structure ultimately allow individual NP MTases to display wide ranging substrate specificity.B ecause of their versatility,N PMTases,w hen used in combination with AdoMet or its artificial analogues, are particularly appealing in the context of biocatalysis and the production of fine chemicals,w here they can help to unlock synthetic routes that would not be accessible using traditional methods.E xcellent overviews of such efforts are provided elsewhere. [2b, 66]

Current and Future Applications
Theever expanding repertoire of doubly-activated cofactor analogues and suitable natural or engineered MTases has provided us with aversatile set of tools for labeling the three major biopolymers. [67] While their most immediate applications might be found in the study of the biological mechanisms underlying epigenetic modification and signaling, it would be hard to overstate the potential uses of such ahighly specific biomolecular toolkit in all areas of research where directed, site-specific labeling is required. In the following paragraphs,aselection of these applications is presented.

Selective Enrichment of Biomolecules
Thespecific enrichment of genetic material is particularly important in the context of high-throughput, high-coverage sequencing efforts. [68] Indeed, even though next-generation sequencing technologies have made whole-genome sequence information relatively easy to obtain, the sheer amount of data produced can be ac onfounding factor.M oreover, diseases such as cancer or viral infections will result in sub populations of cells or even single cells with adistinct genetic makeup against the background of the entire population, giving rise to so-called subclonal genetic heterogeneities. [68b] By labeling target DNA, for example,w ith ab iotincontaining AdoMet analogue,i tc an subsequently be captured, for example by using streptavidin-functionalized particles. [19b] Alternatively,b iomolecules can be labeled using clickable AdoMet analogues and coupled to azide-or alkynefunctionalized beads.This method was recently demonstrated using lambda DNAl abeled with M.TaqIa nd ac ofactor featuring aterminal alkyne. [69] Adifferent study showed DNA capture by labeling with either amine or azide moieties. [50] Subsequently,t he molecules were coupled to the surface of particles via either abiotin functionalized with an NHS ester adduct or biotin modified with adibenzocyclooctyne moiety.
Theu se of MTases in the detection of non-methylated genomic DNAh as been demonstrated by the Klimašauskas group. [50] TheMTase M.SssIwas used to label DNAfeaturing regions of unmethylated CpG sites with acofactor that could be used to biotinylate the DNA. This biotinylated DNAcould subsequently be selectively captured for downstream analysis using microarray screening or sequencing. [50] Next to the identification of genomic regions featuring low methylation, MTase-based labeling strategies have also been applied to the separation of DNAo fi nterest from abackground of other genomic material in the sequencing of the Neanderthal genome.H ere,e nvironmental DNAo f mostly bacterial origin needed to be removed. To achieve this,t he researchers used the fact that mammalian DNAi s frequently methylated at CpG sites,w hich occurs far less frequently in bacterial DNA. Tr eatment with restriction enzymes that specifically target CpG sites largely destroys the bacterial DNA, leaving the endogenous DNAintact. [70] In the referenced application, Mtases are not directly used. Rather,t he DNAo fi nterest (the endogeneous DNA) is methylated in vivo by MTases whereas bacterial DNAisnot. This allows separation of the two through simple restriction endonuclease treatment.
Capturing biomolecules is not just limited to DNA, but can potentially also be used to specifically enrich RNAo r proteins.I ndeed, it has been shown that proteins can be captured by coupling ab iotin tag to proteins with click chemistry. [71] This approach has been used by the Luo group for the specific targeted capturing of proteins using MTasebased labeling (Figure 3). [20a,b,28, 64] Specifically,a nA doMet analogue with alkyne functionality was used for targeted labeling of protein MTase targets. Then, biotin modifed with an azide group was coupled to the proteins using click chemistry,a nd the proteins were subsequently captured using streptavidin beads.T he presence of ac leavable azo linker between the azide and biotin made it possible to separate the captured proteins from the beads after treatment with sodium dithionite (Na 2 S 2 O 4 ). Finally,the strategies presented here might also be useful in the assembly of biomolecular structures [49b] and the positioning of nanoparticles. [49a]

DNA Mapping
Thed iscovery of sequence-specific restriction endonucleases (REases) enabled DNAsequence analysis long before the advent of single-nucleotide sequencing. [72] When aD NA sample is digested by ak nown panel of REases,t he ensuing characteristic fragment lengths can be analyzed using agarose gel electrophoresis,which results in a"DNAfingerprint" with applications in forensics,f or example,. [73] Schwartz et al. further refined the technique by binding linearized DNA molecules to as urface prior to digestion, which results in restriction or sequence "maps" where additional information is contained in the relative location of the different fragments. [74] Site-specific incorporation of fluorescent labels, rather than actual restriction of the DNA, has further contributed to increase the information content of sequence maps. [75] Whereas nicking endonucleases and corresponding fluorescent nucleotide analogues were initially used here, they have since been supplanted by MTase-mediated labeling, which offers an umber of advantages:1 )The use of MTases keeps the DNAb ackbone intact, thereby improving the stability of the DNAm olecules and minimizing unwanted fragmentation. 2) DNAMTases enable amore direct labeling approach compared to nicking endonucleases,w hich increases the efficiencyo ft ransfer.3 )Through careful choice of the DNAM Tases,h igh densities of labeling can be achieved. Indeed, aDNA MTase with a4-base recognition sequence applied to arandom DNAsequence would result in 1label every 256 base pairs (4 4 ).
Thep ast couple of years,s everal approaches to DNA mapping using DNAM Tases have been developed, and excellent reviews on the subject exist. [76] These approaches can be roughly categorized in DNAm apping in nanofluidic channels and DNAm apping using super-resolution microscopy ( Figure 4).
MTase mapping in nanofluidic channels was first demonstrated by the Ebenstein group. [23] Here, phage T7 and phage Lambda DNAw ere labeled using the methyltransferase M.TaqI( recognition sequence:5'-TCGA-3')and asynthetic AdoMet analogue carrying the fluorophore TAMRA. Thel abeled molecules were subsequently stretched in nanochannels,a fter which intensity profiles were extracted using af luorescence microscope.Because of the high density of fluorophores in combination with the use of diffractionlimited microscopy,t he exact location of the labels could not be determined. Instead, the obtained intensity profiles were matched to several computationally generated intensity maps through cross-correlation of the profiles.T hresholding the crosscorrelation scores allows efficient matching of bacteriophages to the correct sequence.M ore recently,t he same group also demonstrated the ability to perform genomic mapping with sub-diffraction-limit resolution in silicon nanochannels, thereby greatly enhancing the information density. [77] Due to the high density of labels,t he work in our laboratory has focused on extracting high-resolution localization information from the DNAm olecules. [19c, 20f] To be able to extract high-resolution localization information from the DNA, it is important that the molecules are fixed on the surface.O ne method for stretching DNAm olecules on as urface is based on flow stretching and attaching the stretched DNAtoasurface coated with poly-l-lysine.Inone such example,b acteriophage T7 (40 kbp) labeled with M.BseCI (recognition sequence:5 '-ATCGAT-3')a nd ab iotin-containing aziridine-modified biotin analogue was coupled to streptavadin-coated quantum dots.T he DNA molecules were subsequently stretched on the surface using capillary flow. [49c] Due to the sparse labeling,the labels could be easily localized. However, the localization suggests there is al arge variation in the positioning owing to deposition inhomogeneity.Analternative method for stretching DNAis based on DNAc ombing,amethod developed by Bensimon et al. in the 90s. [78] With this approach, DNAt ransalkylation using AdoMet analogues carrying either terminal amine or alkyne groups was carried out with the M.HhaIa nd M.TaqI (recognition sequences 5'-TCGA-3' and 5'-GCGC-3')m ethyltransferases.T his gave ah igh density of modified sites, which were subsequently labelled using NHS ester or azide derivatives of the Atto647N dye.F ollowing deposition, subdiffraction-limit imaging was achieved based on stochastic photobleaching and localization of individual emitters. [76a] Figure 3. Use of an AdoMet analogue as areporter of protein methylation. The AdoMet analogue can be utilized by endogenous methyltransferases to label cellular proteins with an alkyne moiety.The modified proteins can then be coupled to afluorescent tag using CuAAC for further characterization. Reproduced from Ref. [21].
This resulted in an approximate resolution of 42 nm (approximately 80 base pairs). [19c] Both these methods suggest ap romising new technique for extracting sequence information from DNAmolecules by studying the position of MTase-mediated labeling on DNA. One example is the typing of bacteriophage DNAmolecules based on ab arcode embedded in the molecule.A nother application could be the detection of copy-number variation, repeats of large genomic elements in the genomes that are hard to find by sequencing. [76a] Finally,t he MTase-based labeling approach can easily be combined with other genomic analysis approaches such as fluorescence in situ hybridization (FISH).

DNA Localization and Spatially Resolved Transcriptomics
Theattachment of reporter groups to biomolecules allows their localization in situ. This has enabled researchers to put detailed functional analysis of biological processes in aspatial context, thereby allowing biological function, that is,m olecular genetics and biochemistry,t ob ec orrelated with information on biological structure (obtained from embryology and histology,f or example). [79] Fluorescent labeling of nucleic acids such as DNAi so ne of the easiest routes to their localization in cells,a nd many methods in fact exist. [80] However,M Tase-mediated transfer of fluorescent groups offers the particular advantage that it allows controlled targeting of the label and its covalent attachment.
In an example from the Weinhold group,C y3 dyes were covalently coupled to pUC19 and pBR322 plasmids using an N-adenosylaziridine cofactor. [51] In this study,t he labeled plasmids were successfully transfected into CHO-k1 cells.Of these transfected cells,2 5% of the cells showed ah igh Cy3 fluorescence intensity in the nucleus,d espite the absence of anuclear import sequence on the plasmids. [51] It can be envisioned that MTase-based labeling strategies could equally contribute to facilitate the in situ study of large scale regulatory networks or efforts in highly multiplexed transcription profiling,asrecently demonstrated Zhuang and co-workers ( Figure 5). [81] Here,the authors used an elaborate library of hybridization probes that were fluorescently labeled and designed to target specific RNAs pecies.B yd esigning multiple,d ifferently labeled probes,a nd using multiple rounds of hybridization, the authors were able to impart au nique color coding to tens to even hundreds of individual RNAs pecies in situ ( Figure 5).
Although this example constitutes ag reat technical and scientific feat, the requirement to design and subsequently use hybridization probes can prove challenging, and indeed, the authors had to apply significant error correction concepts borrowed from computer data encoding to their design of the probes.F urthermore,a lthough this study allowed the simultaneous tracking of large numbers of biomolecular species,it was still limited to RNA. Therefore,t he possibility of achieving of highly directed MTase-mediated labeling of all major classes of biomolecules opens up enticing new prospects to apply massively parallel observation of these species in an effort to truly unravel complex, spatially organized regulatory networks and elucidate cell-to-cell variations in the context of whole tissues. [79]

Epigenetic Analysis
Epigenetic regulation is ac ollective term used to denote the entire spectrum of processes that modulate gene activity in an organism without actually altering the genetic sequence. Cytosine methylation (5mC) and stepwise conversion of the 5mC into hydroxymethyl-(5hmC), formyl-(5fC) and, finally carboxyl-(5caC) cytosine through TET-enzyme-mediated demethylation are common in mammalian cells. [82] These modifications play important roles in embryonic development, stem-cell differentiation, genomic imprinting, neuronal function, and cancer. [83] Additionally,h istones and transcription factors can also be modified through methylation, acylation, or phosphorylation. [84] Because of their role in methylation, MTases,t ogether with suitable AdoMet analogues,are the ideal tools to study these diverse and transient phenomena.

DNA Methylation
Bisulfite conversion is ac ommonly used method for the detection of 5mC,aswell as the other cytosine modifications. It relies on chemical modification of the target species followed by PCR or sequencing-based quantification. [83,85] Unfortunately,t he method suffers from relatively low sensitivity,particularly for low-abundance modifications,aswell as relatively high error rates. [83] Furthermore,t he method only provides ensemble-averaged information, whereas the stochastic nature of DNAmethylation calls for approaches with single-cell resolution. [86] In this context, single-molecule detection methods can offer as olution. In one example,t he methylation status of single DNAmolecules was probed using restriction enzymes for which activity is blocked by the methylated target base.Combining this with optical mapping results in am ethylation map of the DNA. [87] More recently, methyl-CpG-binding proteins were used on DNAstretched in nanochannels, [88] thereby allowing direct detection of the methylation status of single molecules.Amore direct method was used for the detection of hydroxymethylcytosine.H ere, the enzyme T4 b-glucosyltransferase was used to attach glucose molecules with areactive azide moiety to the hydroxymethyl group,w hich could subsequently be coupled with fluorescent dyes. [89] Foramore complete overview of the field, the reader is referred to several excellent reviews. [76] Like restriction enzymes,amethylated base also blocks the activity of methyltransferases.T his characteristic was explored in arecent paper that demonstrated profiling of the unmethylated part of the genome (the "unmethylome"). [50] Unmethylated sites were labeled with the CpG MTase M.SssI and an AdoMet analogue containing either an azide or amine functionality.T he labeled DNAw as then coupled to biotin (biotin functionalized with DBCO or an NHS ester) and extracted using streptavidin-coated microbeads.A fter purification, the DNA was analyzed using microarrays ( Figure 6). This approach suggests ar elatively straight- Figure 5. In the highly multiplexed RNA profiling method of the Zhuang group [81] hundreds of individualm RNA can be correctly be identified and localized in situ.

Angewandte Chemie
Reviews 5196 www.angewandte.org forward way of enriching and subsequently sequencing unmethylated DNA. Because of the sensitivity of the MTase-based method, very low amounts of DNA( 100-300 ng) are needed to complete the analysis.Furthermore,the method was particularly sensitive to regions in the genome with low CpG presence,s omething that presents difficulties for other enrichment techniques such as methylated DNA immuneoprecipitation (MeDIP) and methyl-CpG binding domain capture (MBD). [50] MTase-based labeling of methylated DNAcould easily be combined with single-molecule detection in nanochannels. [76b] Since MTase-based labeling can be targeted towards either adenine or cytosine,t he method could be combined with optical mapping in order to localize regions of (un)methylation. Epigenome mapping could then be performed with adual-color approach, with one color for amap of the DNA, and another color to create am ap of the (un)methylated regions.

Protein Methylation
Thet arget substrates of protein methyltransferases have been extensively studied by the Luo group through the development of an assay termed bioorthogonal profiling of protein methylation (BPPM). Cells are transfected with am utated protein MTase that shows ar educed preference for natural AdoMet compared to the analogue.U pon lysis, the mixture is incubated with ad oubly-activated AdoMet analogue with an azide linker.T arget proteins are then coupled to aclickable dye,for example,dibenzylcyclooctynecoupled dyes,a nd identified by mass spectrometry (MS) analysis or sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE; Figure 3). [21,60,61] Thea pproach was used for detecting the substrates of the human protein MTase PRMT3, which preferably resides within the cytoplasm. [20a] Interestingly,w hile the protein MTase localizes in the cytoplasm, 23 %o ft he modifications were found in the nucleus,thus suggesting abroader role for PRMT3.
More recently,Luo et al. set out to study the methylation status of histones in live cells by hijacking the enzymatic synthesis of AdoMet to create AdoMet analogues instead. A modified version of methionine carrying aclickable group (a terminal alkyne) was transported into the cells and, together with ATP, processed by an engineered MATt of orm the modified cofactor. This modified cofactor was then used with am utant of G9a, which successfully transferred the alkyne group to the histones.T he transfer of the alkyne group to histones was subsequently confirmed using LC-MS. [64] The authors subsequently used this method to attach clickable (azide-conjugated) biotin to the labeled chromatin. This allowed them to enrich the labeled chromatin using streptavidin-coated beads.A fter cleavage of the linker,t he DNA attached to the histones could be purified and sent for sequencing. [64] Indeed, qPCR analysis confirmed the presence of several genes that reside on the targeted histones.Because of the structural similarity between different MTases,t his method is expected to be applicable across aw ide array of protein MTases with varying targets.F or more details on the detection and analysis of protein methylation, the reader is referred to one of several reviews on the topic. [90]

Conclusions and Outlook
Since the conception of labeling biomolecules using MTases,m uch work has been done.T he method offers as imple solution for placing reporter molecules specifically and covalently onto target biomolecules.S of ar, the method has been convincingly shown with more than 20 MTases:9 DNAM Tases,i ncluding both adenine and cytosine DNA MTases,4RNAM Tases,a nd 11 protein MTases.S everal mutated versions of DNAo rp rotein MTases have been developed with improved activity for the various AdoMet analogues,a nd in some cases ap reference of the MTase for the AdoMet analogue over the natural cofactor AdoMet. This preference allows the method to be used in in vivo systems where the natural cofactor AdoMet is still present in the sample mixture.
Tw om ajor groups of AdoMet analogues are currently available for MTase-based labeling.T he aziridine-based cofactors are synthesized chemically and can be used in MTase-mediated labeling reactions in which the whole compound is transferred to the substrate.T he largest group of AdoMet analogues are the double-activated AdoMet analogues,i nw hich the transferrable methyl group of AdoMet is replaced by an unsaturated alkyl group that is used in mTag labeling of biomolecules.T hese analogues are typically synthesized from the precursor AdoHcy. Interestingly,t hese AdoMet analogues can sometimes also be synthesized by hijacking the same enzymes that create the natural cofactor AdoMet in cells,thereby greatly simplifying the synthesis.
This approach for the targeted labeling of biomolecules can subsequently be used for all sorts of applications.T he attachment of biotin or other chemical groups makes it possible to specifically capture low quantities of DNA, mRNA, or proteins.F urthermore,t he method offers aw ay to covalently attach organic fluorescent molecules to target molecules,w hich makes it useful for detection using microscopy-based methods.F inally,t he targeted labeling of DNA offers aw ay to analyze sequence information beyond the standard sequence reads,e ither by imaging far longer DNA molecules or by focusing on the epigenetic code embedded in the DNA.
Thed evelopment of methyltransferases as tools for biotechnology is as tory that showcases the impact that basic chemistry can have in this field. Thee merging applications we have discussed highlight the potential of MTasemediated reactions as ag eneral strategy for targeting and labelling specific sites complex biological samples.T here remains significant work to be done to bring these tools to the non-specialist laboratory,b ut the promise we have outlined should prove as trong driver for future development.