Antibiotic resistance evolved via inactivation of a ribosomal RNA methylating enzyme

Modifications of the bacterial ribosome regulate the function of the ribosome and modulate its susceptibility to antibiotics. By modifying a highly conserved adenosine A2503 in 23S rRNA, methylating enzyme Cfr confers resistance to a range of ribosome-targeting antibiotics. The same adenosine is also methylated by RlmN, an enzyme widely distributed among bacteria. While RlmN modifies C2, Cfr modifies the C8 position of A2503. Shared nucleotide substrate and phylogenetic relationship between RlmN and Cfr prompted us to investigate evolutionary origin of antibiotic resistance in this enzyme family. Using directed evolution of RlmN under antibiotic selection, we obtained RlmN variants that mediate low-level resistance. Surprisingly, these variants confer resistance not through the Cfr-like C8 methylation, but via inhibition of the endogenous RlmN C2 methylation of A2503. Detection of RlmN inactivating mutations in clinical resistance isolates suggests that the mechanism used by the in vitro evolved variants is also relevant in a clinical setting. Additionally, as indicated by a phylogenetic analysis, it appears that Cfr did not diverge from the RlmN family but from another distinct family of predicted radical SAM methylating enzymes whose function remains unknown.     

The Escherichia coli RlmN methyltransferase is a dual-specificity enzyme that modifies both rRNA and tRNA and controls translational accuracy

Modifying RNA enzymes are highly specific for substrate-rRNA or tRNA-and the target position. In Escherichia coli, there are very few multisite acting enzymes, and only one rRNA/tRNA dual-specificity enzyme, pseudouridine synthase RluA, has been identified to date. Among the tRNA-modifying enzymes, the methyltransferase responsible for the m(2)A synthesis at purine 37 in a tRNA set still remains unknown. m(2)A is also present at position 2503 in the peptidyl transferase center of 23S RNA, where it is introduced by RlmN, a radical S-adenosyl-L-methionine (SAM) enzyme. Here, we show that E. coli RlmN is a dual-specificity enzyme that catalyzes methylation of both rRNA and tRNA. The ΔrlmN mutant lacks m(2)A in both RNA types, whereas the expression of recombinant RlmN from a plasmid introduced into this mutant restores tRNA modification. Moreover, RlmN performs m(2)A(37) synthesis in vitro using a tRNA chimera as a substrate. This chimera has also proved useful to characterize some tRNA identity determinants for RlmN and other tRNA modification enzymes. Our data suggest that RlmN works in a late step during tRNA maturation by recognizing a precise 3D structure of tRNA. RlmN inactivation increases the misreading of a UAG stop codon. Since loss of m(2)A(37) from tRNA is expected to produce a hyperaccurate phenotype, we believe that the error-prone phenotype exhibited by the ΔrlmN mutant is due to loss of m(2)A from 23S rRNA and, accordingly, that the m(2)A2503 modification plays a crucial role in the proofreading step occurring at the peptidyl transferase center.     

Insights into the structure,function and evolution of the radical-SAM 23S rRNA methyltransferase Cfr that confers antibiotic resistance in bacteria

The Cfr methyltransferase confers combined resistance to five classes of antibiotics that bind to the peptidyl tranferase center of bacterial ribosomes by catalyzing methylation of the C-8 position of 23S rRNA nucleotide A2503. The same nucleotide is targeted by the housekeeping methyltransferase RlmN that methylates the C-2 position. Database searches with the Cfr sequence have revealed a large group of closely related sequences from all domains of life that contain the conserved CX₃CX₂C motif characteristic of radical S-adenosyl-l-methionine (SAM) enzymes. Phylogenetic analysis of the Cfr/RlmN family suggests that the RlmN subfamily is likely the ancestral form, whereas the Cfr subfamily arose via duplication and horizontal gene transfer. A structural model of Cfr has been calculated and used as a guide for alanine mutagenesis studies that corroborate the model-based predictions of a 4Fe–4S cluster, a SAM molecule coordinated to the iron–sulfur cluster (SAM1) and a SAM molecule that is the putative methyl group donor (SAM2). All mutations at predicted functional sites affect Cfr activity significantly as assayed by antibiotic susceptibility testing and primer extension analysis. The investigation has identified essential amino acids and Cfr variants with altered reaction mechanisms and represents a first step towards understanding the structural basis of Cfr activity.     

Ferredoxins as interchangeable redox components in support of MiaB,a radical S‐adenosylmethionine methylthiotransferase

MiaB is a member of the methylthiotransferase subclass of the radical S‐adenosylmethionine (SAM) superfamily of enzymes, catalyzing the methylthiolation of C2 of adenosines bearing an N⁶‐isopentenyl (i⁶A) group found at position 37 in several tRNAs to afford 2‐methylthio‐N⁶‐(isopentenyl)adenosine (ms²i⁶A). MiaB uses a reduced [4Fe–4S]⁺ cluster to catalyze a reductive cleavage of SAM to generate a 5′‐deoxyadenosyl 5′‐radical (5′‐dA?)—a required intermediate in its reaction—as well as an additional [4Fe–4S]²⁺ auxiliary cluster. In Escherichia coli and many other organisms, re‐reduction of the [4Fe–4S]²⁺ cluster to the [4Fe–4S]⁺ state is accomplished by the flavodoxin reducing system. Most mechanistic studies of MiaBs have been carried out on the enzyme from Thermotoga maritima (Tm), which lacks the flavodoxin reducing system, and which is not activated by E. coli flavodoxin. However, the genome of this organism encodes five ferredoxins (TM0927, TM1175, TM1289, TM1533, and TM1815), each of which might donate the requisite electron to MiaB and perhaps to other radical SAM enzymes. The genes encoding each of these ferredoxins were cloned, and the associated proteins were isolated and shown to support turnover by Tm MiaB. In addition, TM1639, the ferredoxin‐NADP⁺ oxidoreductase subunit α (NfnA) from Tm was overproduced and isolated and shown to provide electrons to the Tm ferredoxins during Tm MiaB turnover. The resulting reactions demonstrate improved coupling between formation of the 5′‐dA? and ms²i⁶A production, indicating that only one hydrogen atom abstraction is required for the reaction.     

4-Hydroxyphenylacetate decarboxylase activating enzyme catalyses a classical S-adenosylmethionine reductive cleavage reaction

4-Hydroxyphenylacetate decarboxylase (4Hpad) is an Fe/S cluster containing glycyl radical enzyme (GRE), which catalyses the last step of tyrosine fermentation in clostridia, generating the bacteriostatic p-cresol. The respective activating enzyme (4Hpad-AE) displays two cysteine-rich motifs in addition to the classical S-adenosylmethionine (SAM) binding cluster (RS cluster) motif. These additional motifs are also present in other glycyl radical activating enzymes (GR-AE) and it has been postulated that these orthologues may use an alternative SAM homolytic cleavage mechanism, generating a putative 3-amino-3-carboxypropyl radical and 5′-deoxy-5′-(methylthio)adenosine but not a 5′-deoxyadenosyl radical and methionine. 4Hpad-AE produced from a codon-optimized synthetic gene binds a maximum of two [4Fe–4S]^2+/+ clusters as revealed by EPR and Mössbauer spectroscopy. The enzyme only catalyses the turnover of SAM under reducing conditions, and the reaction products were identified as 5′-deoxyadenosine (quenched form of 5′-deoxyadenosyl radical) and methionine. We demonstrate that the 5′-deoxyadenosyl radical is the activating agent for 4Hpad through p-cresol formation and correlation between the production of 5′-deoxyadenosine and the generation of glycyl radical in 4Hpad. Therefore, we conclude that 4Hpad-AE catalyses a classical SAM-dependent glycyl radical formation as reported for GR-AE without auxiliary clusters. Our observation casts doubt on the suggestion that GR-AE containing auxiliary clusters catalyse the alternative cleavage reaction detected for glycerol dehydratase activating enzyme.     

The Radical SAM Superfamily

ABSTRACT

The radical S-adenosylmethionine (SAM) superfamily currently comprises more than 2800 proteins with the amino acid sequence motif CxxxCxxC unaccompanied by a fourth conserved cysteine. The charcteristic three-cysteine motif nucleates a [4Fe–4S] cluster, which binds SAM as a ligand to the unique Fe not ligated to a cysteine residue. The members participate in more than 40 distinct biochemical transformations, and most members have not been biochemically characterized. A handful of the members of this superfamily have been purified and at least partially characterized. Significant mechanistic and structural information is available for lysine 2,3-aminomutase, pyruvate formate-lyase, coproporphyrinogen III oxidase, and MoaA required for molybdopterin biosynthesis. Biochemical information is available for spore photoproduct lyase, anaerobic ribonucleotide reductase activation subunit, lipoyl synthase, and MiaB involved in methylthiolation of isopentenyladenine-37 in tRNA. The radical SAM enzymes biochemically characterized to date have in common the cleavage of the [4Fe–4S]^{1 +} –SAM complex to [4Fe–4S]^{2 +}–Met and the 5′ -deoxyadenosyl radical, which abstracts a hydrogen atom from the substrate to initiate a radical mechanism.     

Refolding of a fully functional flavivirus methyltransferase revealed that S‐adenosyl methionine but not S‐adenosyl homocysteine is copurified with flavivirus methyltransferase

Methylation of flavivirus RNA is vital for its stability and translation in the infected host cell. This methylation is mediated by the flavivirus methyltransferase (MTase), which methylates the N7 and 2′‐O positions of the viral RNA cap by using S‐adenosyl‐l ‐methionine (SAM) as a methyl donor. In this report, we demonstrate that SAM, in contrast to the reaction by‐product S‐adenosyl‐l ‐homocysteine, which was assumed previously, is copurified with the Dengue (DNV) and West Nile virus MTases produced in Escherichia coli (E. coli). This endogenous SAM can be removed by denaturation and refolding of the MTase protein. The refolded MTase of DNV serotype 3 (DNV3) displays methylation activity comparable to native enzyme, and its crystal structure at 2.1 Å is almost identical to that of native MTase. We characterized the binding of Sinefungin (SIN), a previously described SAM‐analog inhibitor of MTase function, to the native and refolded DNV3 MTase by isothermal titration calorimetry, and found that SIN binds to refolded MTase with more than 16 times the affinity of SIN binding to the MTase purified natively. Moreover, we show that SAM is also copurified with other flavivirus MTases, indicating that purification by refolding may be a generally applicable tool for studying flavivirus MTase inhibition.     

A thermostable Gm-methylase recognizes the tertiary structure of tRNA

The efficiency of methylation of tRNA by a thermostable tRNA(guanosine-2')-methyltransferase (Gm-methylase) was examined at various temperatures using several species of tRNA isolated from Escherichia coli, yeast and Bacillus subtilis, each possessing different thermal properties. The optimal temperature for the methylation reaction was ca. 20 degrees C lower than the melting temperature of the tRNA in each case. Arrhenius plots of the methylation reactions with various tRNAs gave straight lines below the optimal temperatures in all cases, with similar activation energies of between 10 and 14 kcal/mol. Above the optimal temperatures, the methyl acceptor activity decreased as the incubation temperature was raised to 80 degrees C, at which point the methylase was still active. A correlation was observed between the remaining methyl acceptor activity and the hyperchromicity of tRNA. These results suggest that Gm-methylase recognizes the tertiary structure of tRNA, and it is not the substrate tRNA but the enzyme which is activated by heat.     

Synthesis and Deprotection of Biodegradably and Thermally Protected Dinucleoside‐2′,5′‐Monophosphate Prodrug Model of 2‐5A

Protected dinucleoside‐2′,5′‐monophosphate has been prepared to develop a prodrug strategy for 2‐5A. The removal of enzymatically and thermally labile 4‐(acetylthio)‐2‐(ethoxycarbonyl)‐3‐oxo‐2‐methylbutyl phosphate protecting group and enzymatically labile 3′‐O‐pivaloyloxymethyl group was followed at pH 7.5 and 37 °C by HPLC from the fully protected dimeric adenosine‐2′,5′‐monophosphate 1 used as a model compound for 2‐5A. The desired unprotected 2′,3′‐O‐isopropylideneadenosine‐2′,5′‐monophosphate ( 9 ) was observed to accumulate as a major product. Neither the competitive isomerization of 2′,5′‐ to a 3′,5′‐linkage nor the P–O5′ bond cleavage was detected. The phosphate protecting group was removed faster than the 3′‐O‐protection and, hence, the attack of the neighbouring 3′‐OH on phosphotriester moiety did not take place.     

Unanticipated coordination of tris buffer to the Radical SAM cluster of the RimO methylthiotransferase

Radical SAM enzymes generally contain a [4Fe–4S]^2+/1+ (RS cluster) cluster bound to the protein via the three cysteines of a canonical motif CxxxCxxC. The non-cysteinyl iron is used to coordinate SAM via its amino-carboxylate moiety. The coordination-induced proximity between the cluster acting as an electron donor and the adenosyl–sulfonium bond of SAM allows for the homolytic cleavage of the latter leading to the formation of the reactive 5′-deoxyadenosyl radical used for substrate activation. Most of the structures of Radical SAM enzymes have been obtained in the presence of SAM, and therefore, little is known about the situation when SAM is not present. In this report, we show that RimO, a methylthiotransferase belonging to the radical SAM superfamily, binds a Tris molecule in the absence of SAM leading to specific spectroscopic signatures both in Mössbauer and pulsed EPR spectroscopies. These data provide a cautionary note for researchers who work with coordinative unsaturated iron sulfur clusters.     

Stereochemistry of internucleotidic bond formation by tRNA nucleotidyltransferase from baker's yeast.

Isomer A of adenosine 5'-O-(1-thiotriphosphate) (ATP alpha S) is a substrate for tRNA nucleotidyltransferase from baker's yeast, whereas isomer B is a competitive inhibitor. The tRNA resulting from this reaction has a phosphorothioate instead of a phosphate diester linkage at the last internucleotidic linkage between cytidine and adenosine. On limited digestion of this tRNA with RNase A, one can isolate cytidine 2',3'-cyclic phosphorothioate which can be deaminated to uridine 2',3'-cyclic phosphorothioate. It can be shown that this compound is the endo isomer and that, therefore, the phosphorothioate diester bond in the tRNA must have had the R configuration. This result indicates that no racemization during the condensation of ATP alpha S, isomer A, onto the tRNA had occurred. Whether inversion or retention of configuration had taken place awaits elucidation of the absolute configuration of isomer A of ATP alpha S.     

Radical SAM activation of the B12-independent glycerol dehydratase results in formation of 5'-deoxy-5'-(methylthio)adenosine and not 5'-deoxyadenosine

Activation of glycyl radical enzymes (GREs) by S-adenosylmethonine (AdoMet or SAM)-dependent enzymes has long been shown to proceed via the reductive cleavage of SAM. The AdoMet-dependent (or radical SAM) enzymes catalyze this reaction by using a [4Fe-4S] cluster to reductively cleave AdoMet to form a transient 5'-deoxyadenosyl radical and methionine. This radical is then transferred to the GRE, and methionine and 5'-deoxyadenosine are also formed. In contrast to this paradigm, we demonstrate that generation of a glycyl radical on the B(12)-independent glycerol dehydratase by the glycerol dehydratase activating enzyme results in formation of 5'-deoxy-5'-(methylthio)adenosine and not 5'-deoxyadenosine. This demonstrates for the first time that radical SAM activases are also capable of an alternative cleavage pathway for SAM.     

Generation of adenosyl radical from S-adenosylmethionine (SAM) in biotin synthase

A mechanism of the C―S bond activation of S-adenosylmethionine (SAM) in biotin synthase is discussed from quantum mechanical/molecular mechanical (QM/MM) computations. The active site of the enzyme involves a [4Fe-4S] cluster, which is coordinated to the COO⁻ and NH₂ groups of the methionine moiety of SAM. The unpaired electrons on the iron atoms of the [4Fe-4S]²⁺ cluster are antiferromagnetically coupled, resulting in the S = 0 ground spin state. An electron is transferred from an electron donor to the [4Fe-4S]²⁺-SAM complex to produce the catalytically active [4Fe-4S]⁺ state. The SOMO of the [4Fe-4S]⁺-SAM complex is localized on the [4Fe-4S] moiety and the spin density of the [4Fe-4S] core is calculated to be 0.83. The C―S bond cleavage is associated with the electron transfer from the [4Fe-4S]⁺ cluster to the antibonding σ* C―S orbital. The electron donor and acceptor states are effectively coupled with each other at the transition state for the C―S bond cleavage. The activation barrier is calculated to be 16.0 kcal/mol at the QM (B3LYP/SV(P))/MM (CHARMm) level of theory and the C―S bond activation process is 17.4 kcal/mol exothermic, which is in good agreement with the experimental observation that the C―S bond is irreversibly cleaved in biotin synthase. The sulfur atom of the produced methionine molecule is unlikely to bind to an iron atom of the [4Fe-4S]²⁺ cluster after the C―S bond cleavage from the energetical and structural points of view.     

Dynamics of RNA modification by a multi-site-specific tRNA methyltransferase

In most organisms, the widely conserved 1-methyl-adenosine58 (m¹A₅₈) tRNA modification is catalyzed by an S-adenosyl-L-methionine (SAM)-dependent, site-specific enzyme TrmI. In archaea, TrmI also methylates the adjacent adenine 57, m¹A₅₇ being an obligatory intermediate of 1-methyl-inosine57 formation. To study this multi-site specificity, we used three oligoribonucleotide substrates of Pyrococcus abyssi TrmI (_PabTrmI) containing a fluorescent 2-aminopurine (2-AP) at the two target positions and followed the RNA binding kinetics and methylation reactions by stopped-flow and mass spectrometry. _PabTrmI did not modify 2-AP but methylated the adjacent target adenine. 2-AP seriously impaired the methylation of A₅₇ but not A₅₈, confirming that _PabTrmI methylates efficiently the first adenine of the A₅₇A₅₈A₅₉ sequence. _PabTrmI binding provoked a rapid increase of fluorescence, attributed to base unstacking in the environment of 2-AP. Then, a slow decrease was observed only with 2-AP at position 57 and SAM, suggesting that m¹A₅₈ formation triggers RNA release. A model of the protein–tRNA complex shows both target adenines in proximity of SAM and emphasizes no major tRNA conformational change except base flipping during the reaction. The solvent accessibility of the SAM pocket is not affected by the tRNA, thereby enabling S-adenosyl-L-homocysteine to be replaced by SAM without prior release of monomethylated tRNA.     

Hypothesis of a proton switch in QM/MM modelling of interaction of dUMP analogues with thymidylate synthase

Quantum mechanical, molecular mechanics and molecular dynamics (MD) methods were used to investigate initial steps of 2′-deoxyuridine-5′-monophosphate (dUMP) methylation catalysed by the thymidylate synthase (TS) enzyme. The amino acid residues surrounding the active site within a 10 Å radius sphere were modelled with the combined quantum mechanical (B3LYP/LANL2DZ) and molecular mechanics ONIOM double-layer method. The results indicated the initial nucleophilic attack of Cys146 on dUMP to be concerted with formation of a hydrogen bond to the oxygen O4 of dUMP. Moreover, the proton in the vicinity of the O4 atom appears to act as a ‘proton switch’: if a proton is present near O4, it stabilises the S(Cys146)–C6(dUMP) sulphur–carbon bond, but if it is absent, the sulphur–carbon bond does not form. If the O4 oxygen is replaced by sulphur atom, the ‘switch effect’ does not occur. The suggested correlation between the strength of hydrogen bond involving O4 oxygen and the ability of dUMP to form bonds at C6 corresponds well to the crystal structures of TS complexes available in the Protein Data Bank. In the vast majority of crystal structures, the presence of the S(Cys146)–C6(dUMP) bond was coupled with the presence of hydrogen bond between the dUMP O4 atom and the conserved Asn177. The ‘proton switch’ hypothesis is supported also by the results of MD studies of TS binary complexes, suggesting that average distance separating S(Cys146) and C6(dUMP) becomes distinctly shorter in the presence of hydrogen bonding between Asn177 and O4.