life Review
Homocysteine Editing, Thioester Chemistry, Coenzyme A, and the Origin of Coded Peptide Synthesis † Hieronim Jakubowski 1,2 1
2
†
Department of Microbiology, Biochemistry and Molecular Genetics, New Jersey Medical School, Rutgers University, Newark, NJ 07103, USA;
[email protected] or
[email protected]; Tel.: +1-973-972-8733 Department of Biochemistry and Biotechnology, University of Life Sciences, Poznan 60-632, Poland Presented at the Banbury Center, Cold Spring Harbor Laboratory, NY meeting on “Evolution of the Translational Apparatus and implication for the origin of the Genetic Code”, 13–16 November 2016.
Academic Editor: Koji Tamura Received: 3 January 2017; Accepted: 3 February 2017; Published: 9 February 2017
Abstract: Aminoacyl-tRNA synthetases (AARSs) have evolved “quality control” mechanisms which prevent tRNA aminoacylation with non-protein amino acids, such as homocysteine, homoserine, and ornithine, and thus their access to the Genetic Code. Of the ten AARSs that possess editing function, five edit homocysteine: Class I MetRS, ValRS, IleRS, LeuRS, and Class II LysRS. Studies of their editing function reveal that catalytic modules of these AARSs have a thiol-binding site that confers the ability to catalyze the aminoacylation of coenzyme A, pantetheine, and other thiols. Other AARSs also catalyze aminoacyl-thioester synthesis. Amino acid selectivity of AARSs in the aminoacyl thioesters formation reaction is relaxed, characteristic of primitive amino acid activation systems that may have originated in the Thioester World. With homocysteine and cysteine as thiol substrates, AARSs support peptide bond synthesis. Evolutionary origin of these activities is revealed by genomic comparisons, which show that AARSs are structurally related to proteins involved in coenzyme A/sulfur metabolism and non-coded peptide bond synthesis. These findings suggest that the extant AARSs descended from ancestral forms that were involved in non-coded Thioester-dependent peptide synthesis, functionally similar to the present-day non-ribosomal peptide synthetases. Keywords: aminoacyl-tRNA synthetase; homocysteine editing; thioester; coenzyme A; non-coded peptide synthesis; prebiotic chemistry; thioester world; evolution
1. Introduction Each of the 20 aminoacyl-tRNA synthetases (AARSs) fulfils two important functions in the initial steps in the translation of the Genetic Code: Chemical Activation and Information Transfer. For example, methionyl-tRNA synthetase (MetRS) catalyzes Chemical Activation of the carboxyl group of its cognate amino acid methionine using ATP, which affords Met-AMP bound to the catalytic module of MetRS (Figure 1). The Information Transfer function involves attachment of the activated Met to the 30 adenosine of tRNA(CAU)Met according to the rules of the Genetic Code thereby matching Met with its anticodon CAU, which is read by the AUG codon in the mRNA on the ribosome (Figure 1). AARSs belong to two structurally unrelated classes, of ten enzymes each, which have different catalytic domains indicating their independent evolutionary origin [1,2]. Class I AARSs usually have monomeric structure with a Rossman-fold catalytic domain characterized by the HIGH and KMSKS signature sequences. Class II AARSs have a two- or four-subunit quaternary structure and an antiparallel β sheet catalytic domain with class II-defining motifs. With the exception of LysRS
Life 2017, 7, 6; doi:10.3390/life7010006
www.mdpi.com/journal/life
Life 2017, 7, 6
2 of 26
enzymes, which exist as a Class I or Class II structure in different organisms, other AARSs have a class-specific in the three domains of life. Life 2017, 7,structure 6 2 of 25 Met +
5’ 3’ ATP
AMP + PP
Met 5’ 3’
MetRS
CAU 3’
CAU GUA
5’ mRNA
Figure 1. Chemical activation and information transfer transfer by synthetases (AARSs). Figure 1. Chemical activation and information byaminoacyl-tRNA aminoacyl-tRNA synthetases (AARSs).
AARSs belong to two structurally unrelated classes, of ten enzymes each, which have different
Catalytic domains of Class I and II AARSs exhibit pronounced differences in their modes of catalytic domains indicating their independent evolutionary origin [1,2]. Class I AARSs usually have substrate binding. For example, class I AARSs binddomain ATP incharacterized an extended while class monomeric structure with a Rossman-fold catalytic byconformation, the HIGH and KMSKS II AARSs bind sequences. ATP in a bent with the γ-phosphate folding back over the adenine signature Classconformation II AARSs have a twoor four-subunit quaternary structure and an ring. Whileantiparallel Class I AARSs tRNA via the minor sidemotifs. of its amino acid acceptor β sheetbind catalytic domain with classgroove II-defining With the exception of stem LysRShelix, Class enzymes, II AARSswhich bind tRNA the major groove side. Catalytic domains of Classother I and II AARSs exist asvia a Class I or Class II structure in different organisms, AARSs haveexhibit a class-specific structure Specifically, in the three domains life. such as LysRS, PheRS, HisRS, SerRS, and AspRS, also functional differences. Class IIof AARS, 000 -P Catalytic domains of Class 5I0 ,5 and II 1AARSs exhibit pronounced differences their modes of catalyze formation of diadenosine ,P4 -tetraphosphate (AppppA) [3–5], in a signaling molecule substrate binding. For example, class I AARSs bind ATP in an extended conformation, while class II that participates in transcriptional regulation of IgE-mediated immune response [6,7]. In contrast, AARSs bind ATP in a bent conformation with the γ-phosphate folding back over the adenine ring. Class I AARSs, such as ArgRS and TrpRS, do not possess the AppppA synthetase activity or have While Class I AARSs bind tRNA via the minor groove side of its amino acid acceptor stem helix, >10–100-fold lower activity (ValRS, MetRS, TyrRS) [4,5,8]. Class II AARSs bind tRNA via the major groove side. Catalytic domains of Class I and II AARSs AARSs exhibit high selectivity their cognate acid andastRNA with SerRS, error rates exhibit also functional differences.for Specifically, Classamino II AARS, such LysRS,substrates PheRS, HisRS, −4 to 10−5 and 10−6 , respectively [9–12]. Although in in theand section of amino acids and tRNAs of 10 1 4 AspRS, catalyze formation of diadenosine 5′,5′′′-P ,P -tetraphosphate (AppppA) [3–5], a general unambiguous translation according to the rules of the of Genetic Code is crucialresponse for cellular signaling molecule that participates in transcriptional regulation IgE-mediated immune homeostasis, some species adapted grow and optimally of AppppA ambiguous translation [6,7]. In contrast, Class Ihave AARSs, such astoArgRS TrpRS, in dothe notpresence possess the synthetase or haveconditions, >10–100-fold lowererror activity (ValRS, MetRS, TyrRS) [4,5,8].AARSs achieve unambiguous since,activity under stress higher rates assure survival [13–15]. selectivity for theirtRNAs cognateby amino acid and tRNA substrates with error ratesacids pairing ofAARSs aminoexhibit acidshigh with their cognate preferential binding of cognate amino −4 to 10−5 and 10−6, respectively [9–12]. Although in in the section of amino acids and tRNAs of 10 and a quality control [16] step, in which non-cognate amino acids are selectively edited [10,11,16,17]. general unambiguous translation according to the rules of the Genetic Code is crucial for cellular The quality control step involves either pre-transfer or post-transfer mechanisms, or both [18]. The major homeostasis, some species have adapted to grow optimally in the presence of ambiguous translation pre-transfer mechanism involves hydrolysis of AA~AMP at the catalytic domain, first discovered for since, under stress conditions, higher error rates assure survival [13–15]. AARSs achieve ValRS,unambiguous IleRS, and MetRS while mechanism hydrolysis of AA-tRNA pairing[18,19], of amino acidspost-transfer with their cognate tRNAsinvolves by preferential binding of cognate at a separate editing domain, originally discovered for IleRS [20] and PheRS [21]. Of the 20 extant AARSs, amino acids and a quality control [16] step, in which non-cognate amino acids are selectively edited ten possess an editing function which correctseither errors in aminooracid selection [22]. For some AARS [10,11,16,17]. The quality control step involves pre-transfer post-transfer mechanisms, or both (IleRS,[18]. ValRS, or AlaRS), the editing function is conserved three domains of life, The major pre-transfer mechanism involves hydrolysisthroughout of AA~AMPthe at the catalytic domain, firstwhile discovered and(LeuRS, MetRS [18,19], post-transfer mechanism involves hydrolysis of and editing functionforofValRS, otherIleRS, AARSs ProRS,while or PheRS) is phylogenetically restricted ([23] AA-tRNA at a separate editing domain, originally discovered for IleRS [20] and PheRS [21]. Of the references therein). 20 extantby AARSs, possess an editing function which corrects amino acid selection [22]. Editing AARSsten prevents access of non-proteinogenic aminoerrors acidsinsuch as homocysteine [24–28], For some AARS (IleRS, ValRS, or AlaRS), the editing function is conserved throughout the three ornithine [29], homoserine [10,16], or norvaline [30] to the Genetic Code and effectively partitions domains of life, while editing function of other AARSs (LeuRS, ProRS, or PheRS) is phylogenetically amino acids present in extant organisms into proteinogenic and non-proteinogenic amino acids. Natural restricted ([23] and references therein). non-proteinogenic acids vastly access outnumber the 20 canonical proteinogenic acids found in Editing byamino AARSs prevents of non-proteinogenic amino acids such amino as homocysteine all organisms, plus selenocysteine and[10,16], pyrrolysine encoded some genomes Hundreds [24–28], ornithine [29], homoserine or norvaline [30]intoonly the Genetic Code and[31]. effectively of non-proteinogenic amino acids are known in various species [32]: about 240 in plants, 75 in fungi, partitions amino acids present in extant organisms into proteinogenic and non-proteinogenic amino Natural non-proteinogenic amino acids vastly outnumber the 20 canonical proteinogenic 50 in acids. animals, and 50 in prokaryotes [33]. amino acids found in possess all organisms, plus selenocysteine encoded in only some Of the 10 AARS that the editing function, fiveand editpyrrolysine the thiol amino acid homocysteine [31]. Hundreds of non-proteinogenic acidsIleRS, are known in various species about (Hcy)genomes at the catalytic domain: Class I MetRS,amino LeuRS, ValRS, and Class II [32]: LysRS [10,11]. 240 in plants, 75 in fungi, 50 in animals, and 50 in prokaryotes [33]. Other misactivated amino acids are edited at the catalytic domain, a dedicated editing domain, Of the 10 AARS that possess the editing function, five edit the thiol amino acid homocysteine or both [11,31]. Phylogenetic analyses of structural domains present in proteomes [34] suggest that (Hcy) at the catalytic domain: Class I MetRS, LeuRS, IleRS, ValRS, and Class II LysRS [10,11]. Other catalytic domains of AARSs belong to the oldest fold families and may have appeared about 3.7 billion misactivated amino acids are edited at the catalytic domain, a dedicated editing domain, or both years ago, while separate domains that edit misacylated tRNA appeared later, about 3.2 billion years
Life 2017, 7, 6
3 of 26
ago [35] (these timelines are based on counting relative node, i.e., branch point, distance in the rooted Life 2017, 7, 6 25 trees and using a molecular clock of protein folds to convert relative age into geological3 oftime [36]). Because of Phylogenetic their crucial role in the translation maintenance of the Genetic Code, analysis [11,31]. analyses of structural domains and present in proteomes [34] suggest that catalytic of AARSs belong to the oldest fold families and may have appearedcan about 3.7 billion years into the of AARSs domains structures and mechanisms of reactions catalyzed by AARSs provide insights while separate that edit misacylated tRNA appeared later,examines about 3.2 billion years ago origin andago, evolution of thedomains Genetic Code [35,37]. The present article the links between Hcy [35] (these timelines are based on counting relative node, i.e., branch point, distance in the rooted editing, thioester chemistry, and the origin of the amino acid activation for the coded protein synthesis. trees and using a molecular clock of protein folds to convert relative age into geological time [36]). Available dataBecause suggest thatcrucial ancestral wereand involved in the thiol (coenzyme A, pantetheine) of their role inAARSs the translation maintenance of the Genetic Code, analysis of AARSs reactions structures and mechanisms of reactions catalyzed by AARSs synthesis can providebefore insightsthe intoemergence the aminoacylation and thioester-based non-coded peptide of origin andand evolution of the Genetic Code [35,37]. The present article examines the links between Hcy the Genetic Code the ribosomal protein biosynthetic machinery. 2.
editing, thioester chemistry, and the origin of the amino acid activation for the coded protein synthesis. Available data suggest that ancestral AARSs were involved in the thiol (coenzyme A, Homocysteine (Hcy) is Edited by Class I and Class II Aminoacyl-tRNA Synthetases (AARSs) pantetheine) aminoacylation reactions and thioester-based non-coded peptide synthesis before the of the Genetic Code and the ribosomal protein biosynthetic machinery. One emergence of the selectivity problems in protein biosynthesis is discrimination against the
non-proteinogenic thiol amino acid Hcy, a universal precursor of methionine. Hcy is misactivated 2. Homocysteine (Hcy) is Edited by Class I and Class II Aminoacyl-tRNA Synthetases (AARSs) (Reaction (1)) by Class I MetRS, IleRS, LeuRS [25,26], ValRS [18] and class II LysRS [10,11,29]. One of the selectivity problems in protein biosynthesis is discrimination against the nonproteinogenic thiol amino acid Hcy, a universal precursor of methionine. Hcy is misactivated + IleRS, Hcy + ATP[25,26], AARS • Hcy ∼ class AMP + PPi [10,11,29]. (Reaction I) by ClassAARS I MetRS, LeuRS ValRS [18] and II LysRS AARS + Hcy + ATP ⇄ AARS•Hcy~AMP + PPi
(1)
(I)
Misactivated Hcy is edited by an intramolecular reaction between the side chain thiolate and the Misactivated Hcy is edited by an thioester intramolecular reaction between(Reaction the side chain thiolate and the activated carboxyl of Hcy, affording the Hcy-thiolactone (2)) [18,38]. activated carboxyl of Hcy, affording the thioester Hcy-thiolactone (Reaction II) [18,38].
(II)
(2) Hcy editing does not depend on tRNA [10,18], consumes one mole of ATP per mole prevents attachment Hcy to tRNA, and thus Hcy to the Hcy Hcy-thiolactone editing does[38], notanddepend on tRNAof [10,18], consumes oneaccess mole of Genetic ATP per mole Code. Hcy-thiolactone [38], and prevents attachment of Hcy to tRNA, and thus Hcy access to the The energy of the anhydride bond of Hcy~AMP is conserved in the thioester bond of Genetic Code. Hcy-thiolactone. Consequently, Hcy-thiolactone easily reacts with free amino acids forming Hcy-AA The energy the and anhydride of Hcy~AMP is conserved in the thioester bond of dipeptidesof [39,40] with proteinbond lysine residues forming N-Hcy-protein [39–41]. Hcy-thiolactone. Consequently, Hcy-thiolactone easily reacts with free amino acids forming Hcy-AA Hcy Editing is Universal dipeptides2.1. [39,40] and with protein lysine residues forming N-Hcy-protein [39–41]. Hcy is an important intermediate in the metabolism of Met, Cys, and one-carbon units carried on folates in Bacteria and Eukarya [41,42]. Hcy has also been shown to be an intermediate in Met and 2.1. Hcy Editing is Universal Cys metabolism in Archaea methanogens [43]. Because Hcy is a non-coded amino acid, living
Hcy is an important intermediate in the metabolism Cys,Code. and Indeed, one-carbon units carried organisms must have evolved the ability to prevent its accessof to Met, the Genetic in bacteria (E. coli, M. smegmatis) [24,44], the yeast Saccharomyces cerevisiae [27], plants [45], mice [28], and humans on folates in Bacteria and Eukarya [41,42]. Hcy has also been shown to be an intermediate in Met [28,46], Hcy is edited and metabolized to Hcy-thiolactone by MetRS. In E. coli and S. cerevisiae Hcy and Cys metabolism in Archaea methanogens [43]. Because Hcy is a non-coded amino acid, living thiolactone accumulation is proportional to the expression level of MetRS. In S. cerevisiae, both organismscytoplasmic must have the MetRSs ability edit to prevent access to the Genetic andevolved mitochondrial Hcy [47]. its Editing of endogenous Hcy by Code. MetRS inIndeed, in bacteria (E.cultured coli, M. smegmatis) [24,44], the yeast Saccharomyces cerevisiae with [27],excess plants [45], microbial and mammalian cells is prevented by supplementation Met. In E.mice coli [28], and culturesHcy supplemented two other AARSs LeuRS and IleRS, by in addition humans [28,46], is editedwith andHcy, metabolized to Hcy-thiolactone MetRS.toInMetRS, E. colicatalyze and S. cerevisiae Hcy-thiolactone formation [25,26]. As a result, Hcy-thiolactone formation is fully prevented only by Hcy thiolactone accumulation is proportional to the expression level of MetRS. In S. cerevisiae, both simultaneous supplementation with excess Ile, Leu, and Met. cytoplasmic and MetRSs editgenetic Hcy deficiencies [47]. Editing ofHcy/Cys/Met endogenous Hcy by In allmitochondrial organisms, including human, in the pathways or MetRS in cultured microbial and mammalian cells is prevented by supplementation with excess Met. inadequate supply of cofactors of enzymes participating in Hcy metabolism (folate, In E. coli cobalamin/vitamin B12, Hcy, pyridoxal B6) lead to the accumulation of Hcy and its catalyze cultures supplemented with twophosphate/vitamin other AARSs LeuRS and IleRS, in addition to MetRS, metabolites, including Hcy-thiolactone, which are implicated in cardiovascular and Hcy-thiolactone formation [25,26]. As a result, Hcy-thiolactone formation is fully prevented only by neurodegenerative diseases [41] through mechanisms involving pro-atherogenic changes in gene simultaneous supplementation with excessstructure Ile, Leu, and Met.to amyloid formation [49], activation expression [48], modification of protein [40] leading In all organisms, including human, genetic deficiencies in the Hcy/Cys/Met pathways or inadequate supply of cofactors of enzymes participating in Hcy metabolism (folate, cobalamin/vitamin B12 , pyridoxal phosphate/vitamin B6 ) lead to the accumulation of Hcy and its metabolites, including Hcy-thiolactone, which are implicated in cardiovascular and neurodegenerative diseases [41] through mechanisms involving pro-atherogenic changes in gene expression [48], modification of protein
Life 2017, 7, 6
4 of 26
structure [40] leading to amyloid formation [49], activation of mTORC1 signaling and inhibition of autophagy Life 2017, [50], 7, 6 and induction of inflammatory and autoimmune responses [51–53]. 4 of 25 Structural similarities between Archaeal and Bacterial MetRSs suggest that Hcy can also be edited of mTORC1 signaling inhibition of autophagy [50], inductionPyrococcus of inflammatory and by Archaeal MetRSs. For and example the catalytic domain of and the Archeon abyssi MetRS is autoimmune responses [51–53]. very similar to catalytic domains of E. coli and T. thermophilus MetRSs and can be superimposed Structural similarities between Archaeal and Bacterial MetRSs suggest that Hcy can also be with a root mean square deviation (RMSD) values of 1.7 Å for 481 Cα atoms and 1.6 Å for 406 Cα edited by Archaeal MetRSs. For example the catalytic domain of the Archeon Pyrococcus abyssi MetRS atoms, respectively. Residues important for the synthetic and editing functions of the Bacterial is very similar to catalytic domains of E. coli and T. thermophilus MetRSs and can be superimposed MetRSs conserved in the Archaeal MetRS andof have crystal structures withare a root mean square deviation (RMSD) values 1.7 Åsimilar for 481positions Cα atoms in and 1.6 Å for 406 Cα [54], including a glutamic acid residue, E259, in P. abyssi MetRS homologous to aspartic acid D259 residue atoms, respectively. Residues important for the synthetic and editing functions of the Bacterial in E.MetRSs coli MetRS, which participates as a mechanistic base in the Hcy editing reaction (discussed in are conserved in the Archaeal MetRS and have similar positions in crystal structures [54], including a glutamic acid residue, E259, in P. abyssi MetRS homologous to aspartic acid D259 residue Section 2.2.1 below). in E. coli MetRS, which participates as a mechanistic base in the Hcy editing reaction (discussed in
2.2. Mechanism Hcy Editing Section 2.2.1.ofbelow). HcyMechanism editing isof unique in that it involves an intramolecular Reaction (2), in which the side chain 2.2. Hcy Editing
thiolate of Hcy molecule is a nucleophile, to accomplish editing. Editing reactions of all other amino Hcy editing is unique in that it involves an intramolecular Reaction (II), in which the side chain acids, including a related thio-amino acid Cys, are intermolecular and use water hydroxide as a thiolate of Hcy molecule is a nucleophile, to accomplish editing. Editing reactions of all other amino nucleophile [11]. Hcy is edited by the cyclization to Hcy-thiolactone at the synthetic/editing catalytic acids, including a related thio-amino acid Cys, are intermolecular and use water hydroxide as a site in the Rossman-fold domain of class I AARS [55,56] and at the β catalytic domain of Class II nucleophile [11]. Hcy is edited by the cyclization to Hcy-thiolactone at sheet the synthetic/editing catalytic LysRS [29,38]. site in the Rossman-fold domain of class I AARS [55,56] and at the β sheet catalytic domain of Class II LysRS [29,38].
2.2.1. Methionyl-tRNA Synthetase (MetRS) 2.2.1. Methionyl-tRNA Synthetase (MetRS)
A model for pre-transfer Hcy editing explains how MetRS partitions Met and Hcy between the model for pre-transfer Hcy editing explains MetRS partitions Met and crystal Hcy between the [57], syntheticAand editing pathways, respectively. The how model is supported by the structure synthetic and editing pathways, respectively. The model is supported by the crystal structure [57], structure/function [55,56], and computational [58] studies of E. coli MetRS. In the synthetic pathway, 0 -hydroxyl 0 -terminus structure/function and computational studies ofof E.the coli 3MetRS. In theof synthetic pathway, the activated carboxyl[55,56], of Met reacts with the 2[58] tRNAMet, affording the activated carboxyl of Met reacts with the 2′-hydroxyl of the 3′-terminus of tRNAMet, affording Met-tRNAMet. In the editing pathway, the activated carboxyl of Hcy reacts with the thiolate of its side Met-tRNAMet. In the editing pathway, the activated carboxyl of Hcy reacts with the thiolate of its chain, affording Hcy~ thiolactone. side chain, affording Hcy~thiolactone. Whether an amino acid substrate completes the synthetic or editing pathway is determined by Whether an amino acid substrate completes the synthetic or editing pathway is determined by the partitioning of its side chain specificityand andthiol-binding thiol-binding subsites Met completes the partitioning of its side chainbetween between the the specificity subsites [55].[55]. Met completes the synthetic pathway because is bound boundby bythe thehydrophobic hydrophobic hydrogen-bonding the synthetic pathway becauseits itsside side chain chain is andand hydrogen-bonding interactions with W305 and Y15 specificitysub-site sub-site (Figure interactions with W305 and Y15residues residuesin in the the specificity (Figure 2). 2).
Figure 2. The synthetic/editingactive active site site of synthetase (MetRS): Hydrophobic Figure 2. The synthetic/editing ofE.E.coli colimethionyl-tRNA methionyl-tRNA synthetase (MetRS): Hydrophobic and hydrogen-bonding interactions provide specificity for the cognate substrate L-methionine. and hydrogen-bonding interactions provide specificity for the cognate substrate L-methionine. Superimposition of Cα carbon atoms for the MetRS·Met complex (beige) and free MetRS (light grey), Superimposition of Cα carbon atoms for the MetRS·Met complex (beige) and free MetRS (light grey), solved at 1.8 Å resolution, shows movements of active site residues upon binding of Met. Residue solved at 1.8 Å resolution, shows movements of active site residues upon binding of Met. Residue colors are red in the MetRS·Met complex and green in free MetRS, and L-methionine is magenta. colors are redwith in the MetRS·from Met complex Reprinted permission ref. [57]. and green in free MetRS, and L-methionine is magenta. Reprinted with permission from reference [57].
Life 2017, 7, 6 Life 2017, 7, 6 Life 2017, 7, 6
5 of 26 5 of 25 5 of 25
In contrast, contrast, the side chain of Hcy, Hcy, missing missing the the methyl methyl group group of of Met, Met, interacts interacts weakly weakly with with the the In contrast, theThis side allows chain of Hcy, methyl groupwith of Met, weakly with the specificity sub-site. the sidemissing chain ofthe Hcy to interact D259interacts [58] in the thiol-binding [58] in the thiol-binding specificity sub-site. allowsediting the side of Hcyto toHcy-thiolactone interact with D259 [58] in thiol-binding sub-site [55], which facilitates by cyclization (Figure 3). Consistent with [55], whichThis facilitates editing bychain cyclization to Hcy-thiolactone (Figure 3).the Consistent with sub-site [55], which facilitates editing by cyclization to Hcy-thiolactone (Figure 3). Consistent with this model, mutations of W305 and Y15 residues, which form the hydrophobic Met-binding sub-site, this model, mutationsdiscrimination of W305 and Y15 residues, which reduce the discrimination by the the enzyme [56]. form the hydrophobic Met-binding sub-site, reduce the Hcy/Met Hcy/Met by enzyme reduce the Hcy/Met discrimination by the enzyme [56].
Figure 3. 3. Editing Editing of of miscativated miscativated homocysteine homocysteine (Hcy~AMP) (Hcy~AMP) at at the the catalytic catalytic module module of of an an AARS: AARS: The The Figure Figure 3. Editing of miscativated homocysteine adenylate (Hcy~AMP) at the catalytic module and of anAMP, AARS: The MetRS-catalyzed cyclization of homocysteinyl to form Hcy-thiolactone which MetRS-catalyzed cyclization of homocysteinyl adenylate to form Hcy-thiolactone and AMP, which are MetRS-catalyzed cyclization of homocysteinyl adenylate to form Hcy-thiolactone and AMP, which are subsequently released the synthetic/editing active of MetRS. subsequently released fromfrom the synthetic/editing active site site of MetRS. are subsequently released from the synthetic/editing active site of MetRS.
Computational studies [58] suggest that D259 plays an essential role as a mechanistic base that studies [58] suggest that D259 plays an essential role as a mechanistic base that Computational studies deprotonates the side chain thiol in Hcy~AMP at the catalytic module of MetRS (Figure 4). In the deprotonates the side chain thiol in in Hcy~AMP at thethe catalytic module of MetRS (Figure 4). In 4). the In initial deprotonates the side chain thiol Hcy~AMP module of MetRS (Figure the initial MetRS·Hcy~AMP complex the distance at SHcy···Ocatalytic Asp259 (5.30 Å ) is markedly shorter than the MetRS·MetRS· Hcy~AMP complex the distance SHcy ···OSAsp259 (5.30 is markedly shorter than the than distance initial Hcy~AMP complex the distance Hcy···O Asp259 Å) (5.30 Å ) is markedly shorter the distance SHcy···Ophos in Hcy~AMP (7.15 Å ) (Table 1). The rate-limiting step in Hcy-thiolactone SHcy ···Ophos Hcy~AMP (7.15 Å) (7.15 (TableÅ 1). The rate-limiting step−1 in Hcy-thiolactone formation distance S Hcyin · · · O phos in Hcy~AMP ) (Table 1). The rate-limiting step in Hcy-thiolactone formation is the rotation about the Cβ−Cγ bond with 27.5 kJ·mol energy barrier. This is more −1 energy is the rotation about the Cabout −Cγthe bond with 27.5 kJ ·mol27.5 is more formation is the rotation C β−C γ bond with kJ· mol−1barrier. energyThis barrier. Thisfavorable is which more β −1 favorable than 98.2−51 kJ·mol energy barrier for an alternative substrate-assisted mechanism in −1 than 98.25than kJ·mol energy barrier for an for alternative substrate-assisted mechanism in which the favorable 98.2 5 kJ· mol energy barrier an alternative substrate-assisted mechanism in which the non-bridging oxygen of phosphate in Hcy~AMP acts as a base. non-bridging oxygen of phosphate in Hcy~AMP actsacts as aas base. the non-bridging oxygen of phosphate in Hcy~AMP a base.
Figure 4. Energetics of Hcy~AMP editing at the catalytic module of MetRS. Asp259 is a mechanistic Figure 4. Energetics of Hcy~AMP editing at catalytic Reproduced module of MetRS. Asp259 is afrom mechanistic base that deprotonates the side chain thiol inthe Hcy~AMP. with permission ref. base [58]. Figure 4. Energetics of Hcy~AMP editing at the catalytic module of MetRS. Asp259 is a mechanistic base that deprotonates the side chain thiol in Hcy~AMP. Reproduced with permission from ref. [58]. that deprotonates the side chain thiol in Hcy~AMP. Reproduced with permission from reference [58].
Life 2017, 7, 6
6 of 26
Life 2017, 7, 6
6 of 25
Table 1. Average distances calculated from molecular dynamics simulation for Hcy~AMP bound in Table 1. Average distances calculated from molecular dynamics simulation for Hcy~AMP bound in the active site of Class I AARSs. the active site of Class I AARSs. Average Distance,ÅÅ Average Distance, MetRS LeuRS ValRS ValRS IleRS IleRS MetRS LeuRS SHcy ···OSAsp/Glu Hcy···OAsp/Glu 5.30 5.51 4.944.94 4.70 4.70 5.30 5.51 SHcy ···C 3.91 5.19 Scarb Hcy···Ccarb 3.91 5.19 4.144.14 4.66 4.66 Ccarb ···O 4.25 4.39 Asp/Glu Ccarb···OAsp/Glu 4.25 4.39 5.265.26 4.10 4.10 SHcy ···Ophos 6.99 7.78 7.01 4.25 SHcy···Ophos 6.99 7.78 7.01 4.25
Cognate Met can also enter the editing pathway when the thiol-binding subsite subsite is occupied occupied by a thiol ofof Hcy [55]. Under these circumstances the activated carboxyl and thiol thiol mimicking mimickingthe theside sidechain chain Hcy [55]. Under these circumstances the activated carboxyl and functions are on are separate molecules and MetRS Met:thiolaligase that catalyzes synthesis of thiol functions on separate molecules andbecomes MetRS abecomes Met:thiol ligase that catalyzes Met thioesters (Figure 5). (Figure 5). synthesis of Met thioesters
activity of MetRS. MetRS. When When the active site is occupied Figure 5. Amino acid:coenzyme A (CoA-SH) ligase activity by Met-tRNA Met-tRNAoror Met-AMP, the thiol subsite is occupied by CoA-SH, MetRS the catalyzes the by Met-AMP, andand the thiol subsite is occupied by CoA-SH, MetRS catalyzes formation formation of Met-S-CoA thioester. An R represents the bulk of CoA-SH or other thiol molecule. of Met-S-CoA thioester. An R represents the bulk of CoA-SH or other thiol molecule.
2.2.2. Leucyl-tRNA Leucyl-tRNA Synthetase Synthetase (LeuRS), (LeuRS), Isoleucyl-tRNA Isoleucyl-tRNA Synthetase Synthetase (IleRS), (IleRS), Valyl-tRNA Valyl-tRNA Synthetase 2.2.2. (ValRS) Synthetase (ValRS) A similar model explains Hcy editing [25,26] by related Class I AARS. ValRS, LeuRS, and IleRS have active activesites, sites,D490, D490, E532, E550, respectively, that correspond ofand MetRS and are have E532, andand E550, respectively, that correspond to D359toofD359 MetRS are similarly similarly positioned with tocenter the Ccarb center of the substrate Hcy~AMP [58]. Computational positioned with respect to respect the Ccarb of the substrate Hcy~AMP [58]. Computational analyses analyses that the Hcy~AMP in theirsites active in aconformation linear conformation show thatshow the Hcy~AMP substratesubstrate binds in binds their active in asites linear similar similar to that to that observed for MetRS. The average C carb · · · O Asp/Glu distances are within 1.16 Å of each other observed for MetRS. The average Ccarb ···OAsp/Glu distances are within 1.16 Å of each other whereas all all distances SHcy···Ccarbare distances are within 1.24 Å of(Table each 1). other (Table 1). Moreforimportantly, for Swhereas within 1.24 Å of each other More importantly, MetRS, LeuRS, Hcy ···Ccarb MetRS, LeuRS, and ValRS, Hcy···OAsp/Glu distance is significantly shorter thanSthe and ValRS, the average SHcythe ···Oaverage distance is significantly shorter than the average ···Ophos Hcy average Asp/Glu S S Hcy · · · O phos distance by 2.15 Å , 2.27 Å , and 2.06 Å , respectively. Thus, each of these Asp/Glu residues distance by 2.15 Å, 2.27 Å, and 2.06 Å, respectively. Thus, each of these Asp/Glu residues can act as a can act as abase mechanistic base that the deprotonates the side chain thiol in facilitate Hcy~AMP and facilitate mechanistic that deprotonates side chain thiol in Hcy~AMP and Hcy-thiolactone Hcy-thiolactone formation MetRS, and itValRS. it is Glu550 less clear formation by MetRS, LeuRS,byand ValRS.LeuRS, However, is less However, clear whether canwhether act as a Glu550 base in can act as a base formation in Hcy-thiolactone Hcy-thiolactone catalyzedformation by IleRS. catalyzed by IleRS. 2.2.3. Lysyl-tRNA Lysyl-tRNA Synthetase Synthetase (LysRS) (LysRS)Edits EditsHomocysteine Homocysteine(Hcy), (Hcy),Ornithine Ornithine(Orn), (Orn),Homoserine Homoserine(Hse), but Mischarges tRNALys withLysProteinogenic AminoAmino Acids Acids (Hse), but Mischarges tRNA with Proteinogenic Hcy editing occurs also at the β β sheet catalytic domain of Class II E. E. coli coli LysRS [38], which which also Lys Lys Lys Lys edits Orn and Hse. Editing by LysRS is not affected by tRNA and there is no tRNA mischarging Orn and Hse. Editing by LysRS is not affected by tRNA and there is no tRNA Lys with several other amino acids (Arg, Thr, with with Hcy Hcy or or Orn [29]. However, LysRS LysRS mischarges mischarges tRNA tRNALys with several other amino acids (Arg, Thr, Lys , Thr-tRNA Lys , and Lys, Thr-tRNA Lys, Met, Cys, Ser) andand doesdoes not deacylate mischarged Arg-tRNA Cys, Leu, Leu,Ala, Ala,oror Ser) not deacylate mischarged Arg-tRNA Lys Lys Lys Lys Met-tRNA . Recent data show that Met-tRNA (and other tRNAs mischarged with Met) are and Met-tRNA . Recent data show that Met-tRNA (and other tRNAs mischarged with Met) formed in E. E. coli coli and and mammalian mammalian cells cells in in response response to to stress stress conditions conditions[13,14]. [13,14].
Life 2017, 7, 6
7 of 26
Life 2017, 7, 6
7 of 25
3. Expanding Expanding the the Genetic Genetic Code: Code: Decoding Decoding Methionine Methionine Codons Codons by by Homocysteine Homocysteine 3. By preventing preventing the Hcy to tRNA, the the editing editing reaction reaction assures assures that excluded By the attachment attachment of of Hcy to tRNA, that Hcy Hcy is is excluded from the Genetic Code. Weak interactions of the side chain of Hcy with the specificity subsite allow from the Genetic Code. Weak interactions of the side chain of Hcy with the specificity subsite allow binding of of the the Hcy Hcy side side chain chain to to the the thiol-binding thiol-binding subsite subsite in in the the catalytic catalytic domain domain of of MetRS MetRS [55,56]. [55,56]. binding Thus, modifications of the side chain of Hcy that increase binding to the specificity subsite should Thus, modifications of the side chain of Hcy that increase binding to the specificity subsite should Met This is achieved by S-nitrosylation of Hcy to prevent prevent editing editing and and facilitate facilitate the the transfer transfer to to tRNA tRNAMet.. This is achieved by S-nitrosylation of Hcy to S-nitroso-Hcy (S-NO-Hcy), which binds to the MetRS with affinity10-fold 10-foldgreater greater than Hcy and S-nitroso-Hcy (S-NO-Hcy), which binds to the MetRS with affinity than Hcy and is is activated by MetRS to form S-NO-Hcy~AMP [59]. In contrast to Hcy-AMP which is edited, activated by MetRS to form S-NO-Hcy~AMP [59]. In contrast to Hcy-AMP which is edited, S-NO-Hcy~AMP subsite which leads to S-NO-Hcy~AMP is is resistant resistantto toediting editingdue duetotostronger strongerbinding bindingtotothe thespecificity specificity subsite which leads Met Met thethe transfer of S-NO-Hcy to tRNA , affording S-NO-Hcy~tRNA to transfer of S-NO-Hcy to tRNAMet , affording S-NO-Hcy~tRNAMet(Figure (Figure6). 6).
Figure by MetRS MetRS [59]. [59]. Figure 6. 6. Aminoacylation Aminoacylation of of tRNA tRNA with with S-NO-Hcy S-NO-Hcy catalyzed catalyzed by Met has a similar susceptibility to spontaneous deacylation as Met~tRNAMet The S-NO-Hcy-tRNAMet Met The S-NO-Hcy-tRNA has a similar susceptibility to spontaneous deacylation as Met~tRNA Met affords Met, the least with a half-life of 28 min. De-nitrosylation of S-nitroso-Hcy~tRNAMet Hcy~tRNAMet with a half-life of 28 min. De-nitrosylation of S-nitroso-Hcy~tRNA affords Hcy~tRNA , the least stable aminoacyl-tRNA known, which spontaneously deacylates with a half-life of 15 s to form stable aminoacyl-tRNA known, which spontaneously deacylates with a half-life of 15 s to form Met. Hcy-thiolactone and free tRNAMet Hcy-thiolactone and free tRNA . Met As expected, S-NO-Hcy~tRNAMet is a substrate for protein synthesis on ribosomes, which allows As expected, S-NO-Hcy~tRNA is a substrate for protein synthesis on ribosomes, which allows translational incorporation of S-NO-Hcy into protein at positions normally occupied by Met [59]. translational incorporation of S-NO-Hcy into protein at positions normally occupied by Met [59]. For example, when cultures of E. coli metE mutant cells (unable to metabolize Hcy to Met) expressing For example, when cultures of E. coli metE mutant cells (unable to metabolize Hcy to Met) expressing mouse dihydrofolate reductase (DHFR) protein were supplemented with S-NO-Hcy, the DHFR mouse dihydrofolate reductase (DHFR) protein were supplemented with S-NO-Hcy, the DHFR protein protein was found to contain Hcy. Control experiments in which E. coli metE cultures were was found to contain Hcy. Control experiments in which E. coli metE cultures were supplemented with supplemented with Hcy or Hcy-thiolactone, instead of S-NO-Hcy, show that there is no incorporation Hcy or Hcy-thiolactone, instead of S-NO-Hcy, show that there is no incorporation of Hcy into bacterial of Hcy into bacterial proteins [59]. Globin and luciferase, produced in an in vitro mRNA-programmed proteins [59]. Globin and luciferase, produced in an in vitro mRNA-programmed rabbit reticulocyte rabbit reticulocyte protein synthesis system supplemented with S-NO-Hcy~tRNAMet contain Hcy at Met protein synthesis system supplemented with S-NO-Hcy~tRNA contain Hcy at positions normally positions normally occupied by Met. occupied by Met. Translationally incorporated Hcy has also been identified in protein from cultured human Translationally incorporated Hcy has also been identified in protein from cultured human vascular vascular endothelial cells (HUVECs) (Table 2), which endogenously produce nitric oxide and Sendothelial cells (HUVECs) (Table 2), which endogenously produce nitric oxide and S-nitroso-Hcy [60]. nitroso-Hcy [60]. Translationally incorporated Hcy is resistant to Edman degradation whereas postTranslationally incorporated Hcy is resistant to Edman degradation whereas post-translationally translationally incorporated Hcy (by the reaction of Hcy-thiolactone with protein lysine residues [40]) incorporated Hcy (by the reaction of Hcy-thiolactone with protein lysine residues [40]) is not (Table 2, is not (Table 2, last row) last, which allows to distinguish between the two mechanisms [61]. last row) last, which allows to distinguish between the two mechanisms [61]. Taken together, these findings show that Hcy can gain an access to the Genetic Code by nitric oxide-mediated invasion of the methionine-coding pathway [59].
Table 2. Translational and post-translational incorporation of Hcy into HUVEC protein [60,62].
Table 2. Translational and post-translational incorporation of Hcy into HUVEC protein [60,62]. Translational Post-Translational Cell Labeling Conditions 35 S]Hcy-Lys-Protein Cell labeling Conditions Translational Post-translational [35 S]Hcy-Protein [35 S]Met-Protein εN-[ [35S]Hcy-protein [35S]Met-protein N-[35S]Hcy-Lys-protein % % % % % 38 37% 25 [35 S]Hcy (10 µM, 50 µCi/mL) 35S]Hcy (10 M, 50 Ci/mL) 37 25 38
