Enzymatic reconstitution of ribosomal peptide backbone thioamidation

Edited by Jerrold Meinwald, Cornell University, Ithaca, NY, and approved February 13, 2018 (received for review December 21, 2017)
March 5, 2018
115 (12) 3030-3035


Thioamidation as a posttranslational modification is exceptionally rare, with only one protein example known (methyl-coenzyme M reductase, MCR), as well as a few ribosomal natural products. The genes involved in MCR thioamidation have recently been elucidated, but the enzymes have yet to be characterized. Herein, we report the in vitro reconstitution and substrate preferences of peptidic thioamidation using peptides and enzymes derived from methanogenic archaea. We demonstrate that MCR thioamidation requires an ATP-dependent YcaO enzyme and a sulfide source. Our results shed light on the biosynthesis of other thioamide-containing compounds, which bioinformatics surveys predict to be considerably more numerous than currently appreciated, thus laying a foundation for assigning biological functions for this posttranslational modification.


Methyl-coenzyme M reductase (MCR) is an essential enzyme found strictly in methanogenic and methanotrophic archaea. MCR catalyzes a reversible reaction involved in the production and consumption of the potent greenhouse gas methane. The α-subunit of this enzyme (McrA) contains several unusual posttranslational modifications, including the only known naturally occurring example of protein thioamidation. We have recently demonstrated by genetic deletion and mass spectrometry that the tfuA and ycaO genes of Methanosarcina acetivorans are involved in thioamidation of Gly465 in the MCR active site. Modification to thioGly has been postulated to stabilize the active site structure of MCR. Herein, we report the in vitro reconstitution of ribosomal peptide thioamidation using heterologously expressed and purified YcaO and TfuA proteins from M. acetivorans. Like other reported YcaO proteins, this reaction is ATP-dependent but requires an external sulfide source. We also reconstitute the thioamidation activity of two TfuA-independent YcaOs from the hyperthermophilic methanogenic archaea Methanopyrus kandleri and Methanocaldococcus jannaschii. Using these proteins, we demonstrate the basis for substrate recognition and regioselectivity of thioamide formation based on extensive mutagenesis, biochemical, and binding studies. Finally, we report nucleotide-free and nucleotide-bound crystal structures for the YcaO proteins from M. kandleri. Sequence and structure-guided mutagenesis with subsequent biochemical evaluation have allowed us to assign roles for residues involved in thioamidation and confirm that the reaction proceeds via backbone O-phosphorylation. These data assign a new biochemical reaction to the YcaO superfamily and paves the way for further characterization of additional peptide backbone posttranslational modifications.
Methyl-coenzyme M reductase (MCR) catalyzes the reversible conversion of methyl-coenzyme M (2-methylmercaptoethanesulfonate, CoM) and coenzyme B (7-thioheptanoylthreonine-phosphate, CoB) to methane and CoB-CoM heterodisulfide and is found only in anaerobic archaea (1, 2). This reaction plays a key role in the global carbon cycle by maintaining steady-state levels of atmospheric methane, a potent greenhouse gas (3). MCR exhibits a hexameric (α2β2γ2) architecture (2) with an essential nickel porphinoid coenzyme F430 present in the active site (47). One remarkable feature of MCR is the posttranslational modification (PTM) of several active site residues in the α-subunit (McrA). These include thioglycine, 3-methylhistidine, S-methylcysteine, 5-methylarginine, 2-methylglutamine, and didehydroaspartate (810). Among these, thioglycine, 5-methylarginine, and 3-methylhistidine are present in all methanogens examined thus far (9). The functions of these PTMs have yet to be elucidated although recent work has suggested a potential structural role for thioglycine (11).
There are very few naturally occurring thioamides, and most known examples are bacterial natural products (12, 13). These include ribosomally synthesized and posttranslationally modified peptide (RiPP) natural products (14) such as the thioviridamides (1517), methanobactins (18, 19), thioamide-containing thiopeptides (20, 21), as well as the non-RiPP, peptide-like compound closthioamide (SI Appendix, Fig. S1) (22). Upon inspection of the thioviridamide biosynthetic gene cluster (23), it was hypothesized that the YcaO and TfuA proteins would be involved in thioamidation; however, this proposal awaits biochemical validation. Nevertheless, taken with our earlier work on azoline-forming YcaOs (12, 24, 25), the recent genetic deletion studies in Methanosarcina acetivorans have shown that ycaO and tfuA are required for the thioamidation of McrA-Gly465 (11).
YcaO enzymes catalyze the ATP-dependent backbone cyclodehydration of Cys, Ser, and Thr to the corresponding thiazoline and (methyl)oxazoline heterocycle (Fig. 1). This modification is found in several RiPP classes, including the linear azol(in)e-containing peptides, thiopeptides, and certain cyanobactins (12). Well-characterized YcaOs require a partner protein for efficient azoline formation, which is found either fused as a single polypeptide with the YcaO or encoded adjacently as a separate protein. Two varieties of partner proteins are known, with one type resembling E1-ubiquitin–activating enzymes and the other referred to as “ocin-ThiF” proteins (12, 26, 27). These partner proteins harbor an ∼90-residue N-terminal RiPP precursor peptide recognition element (RRE) (28), which recognizes and binds a particular region of the precursor peptide. With few exceptions (29, 30), the recognition sequence resides in the N-terminal “leader” region of the precursor peptide. Once bound, the YcaO protein then enzymatically modifies the C-terminal “core” region. Bioinformatics surveys predict that thousands of YcaO proteins will depend on a partner protein; however, such hypotheses await experimental validation (31). In contrast, two recently characterized YcaOs from bottromycin biosynthesis are shown to catalyze the formation of thiazoline and macrolactamidine without a partner (29, 30).
Fig. 1.
Comparison of reactions catalyzed by YcaO enzymes. (Upper) Biochemically characterized azoline-forming YcaO proteins catalyze the ATP-dependent cyclodehydration of cysteine, serine, and threonine residues. Shown is the conversion of peptidic cysteine to thiazoline, which proceeds via an O-phosphorylated hemiorthoamide intermediate. (Lower) An analogous reaction is proposed for the thioamide-forming YcaOs, which utilize an external source of sulfide (such as Na2S), as opposed to an adjacent cysteine, as the nucleophile.
Based on the multitude of additional examples where ycaO and tfuA are adjacently encoded (16, 23) and the results of our studies in M. acetivorans, which show that genetic deletion of ycaO and tfuA abolished thioglycine formation (11), we proposed that TfuA may partner with the YcaO protein to enhance peptidic thioamidation (Fig. 1). Plausible roles for TfuA include binding the peptide substrate, regulating the usage of ATP by the YcaO, or assisting in the delivery of sulfide equivalents. Bioinformatic survey reveals that certain methanogens lack tfuA in their genome (SI Appendix, Fig. S2) (11). Whether these YcaOs can function alone or require a different partner protein remains unknown.
In this work, we confirmed in vitro the role of YcaO proteins in peptidic thioamidation. Using heterologously expressed and purified YcaO and TfuA proteins from M. acetivorans and McrA peptide fragments, thioamidation was successfully reconstituted and found to be ATP- and sulfide-dependent. Mutational analysis of the substrate identified critical residues flanking Gly465. The activities of two TfuA-independent YcaOs from the hyperthermophilic methanogens Methanopyrus kandleri and Methanocaldococcus jannaschii were also reconstituted. The crystal structures of these proteins were solved, supporting our mutational analysis of residues involved in ATP binding. Isotope-labeling studies and product analysis further demonstrated that thioamide-forming YcaOs utilize a backbone phosphorylation mechanism akin to azoline-forming YcaOs (12). Our data on posttranslational thioamidation demonstrate that the modification is installed by a mechanism distinct from nucleoside thioamidation (3234).


Bioinformatics Survey of YcaOs from Methanogenic Archaea.

All available YcaO sequences from methanogens were retrieved from UniProt (35) (n = 154 as of mid-2017). All methanogenic archaea encode at least one YcaO protein. The frequency of cooccurring genes was next analyzed with ∼70% of ycaO genes being encoded adjacently to tfuA. An additional ∼20% of methanogenic archaea distally encode TfuA. The remaining ∼10% do not contain a tfuA homolog identifiable by BLASTp search, as previously reported (11). We then created a sequence similarity network (SSN) and maximum likelihood tree from the YcaO sequences (SI Appendix, Fig. S2) to categorize the sequences based on their relationship with TfuA. The YcaO sequences from methanogens that adjacently encode ycaO/tfuA were more sequence similar to each other than the YcaO sequences from methanogens that encode ycaO distally to tfuA or lack the tfuA gene altogether (SI Appendix, Fig. S2). On the tree, there are two distinct clades comprised of a mixture of tfuA-distal and tfuA-independent YcaOs, which suggests that the loss of tfuA from these strains may have occurred independently. To reveal the important residues for YcaO function, five YcaO sequences were chosen from five separate clades of the phylogenetic tree to generate a diversity-oriented multiple sequence alignment (SI Appendix, Fig. S3). It is evident that the ATP- and Mg2+-coordinating residues (e.g., Glu78, Glu81, Glu164, and Glu167 in M. kandleri YcaO) are highly conserved, as previously reported for azoline-forming YcaOs (25).

In Vitro Reconstitution of Thioglycine Formation.

To determine if YcaO and TfuA directly perform McrA thioamidation, as suggested by an earlier study (11), we prepared maltose-binding protein (MBP) fusions of the YcaO, TfuA, and McrA proteins (SI Appendix) from M. acetivorans. While the YcaO and TfuA proteins readily purified, McrA was severely degraded and, hence, unusable (SI Appendix, Fig. S4). Given the buried location of the thioamide in intact MCR (2), we hypothesized that thioamidation would likely occur on an unstructured, linear McrA substrate before complexation with the other MCR subunits. We thus prepared McrA fragments centering on Gly465 to be tested as substrates by M. acetivorans TfuA and YcaO. The substrates tested were the 11-, 21-, and 51-residue peptides encompassing McrA positions Arg460-Asp470, Lys455-Thr475, and Ala440-Leu490, respectively (SI Appendix, Figs. S4 and S5). We first treated the 51-mer substrate with YcaO and TfuA in the presence of Na2S and ATP. The reaction mixture was then analyzed by matrix-assisted laser desorption/ionization-time-of-flight mass spectrometry (MALDI-TOF-MS), which showed a [M + 16] ion as expected for thioamidation (SI Appendix, Fig. S6). Omission of any reaction component resulted in no observable reaction. The shorter peptides were also accepted as substrates by M. acetivorans YcaO/TfuA to varying efficiencies with the 11-mer being completely converted to the [M + 16] species (Fig. 2 and SI Appendix, Figs. S7 and S8). Analysis by high-resolution electrospray ionization mass spectrometry (HR-ESI-MS) and the requirement for an external sulfide source support thioamidation (formal replacement of oxygen by sulfur, +15.9772 Da, error ∼0.3 ppm), as opposed to oxidation or hydroxylation (formal addition of oxygen, +15.9949 Da, error >12 ppm), which are not known to be catalyzed by YcaO proteins (SI Appendix, Fig. S9). Tandem HR-ESI-MS localized the site of thioamide installation to the position equivalent to Gly465, the native modification site in McrA (SI Appendix, Fig. S9). In addition to using a chemical source of sulfide, thioamidation was also observed upon using the cysteine desulfurase from M. acetivorans (IscS) with the reactions supplemented with Cys and pyridoxal phosphate (Fig. 2).
Fig. 2.
MALDI-TOF-MS analysis of McrA 11-mer peptide thioamidation using M. acetivorans YcaO-TfuA. The McrA fragment was from M. acetivorans with the residue naturally thioamidated in red. (A) Mass spectrum of the unmodified peptide, m/z 1,428 Da. (B) Mass spectrum of the peptide after reaction with the full array of reactants. (C) Identical to B except an enzymatic sulfide source was used rather than a chemical donor (PLP, pyridoxal phosphate). (D) Identical to B but with omission of YcaO. (E) Omission of TfuA. (F) Omission of sodium sulfide.

Substrate Scope and Site Selectivity of Thioamidation.

To further investigate the substrate tolerance of the M. acetivorans YcaO/TfuA proteins, we prepared a panel of Ala-substituted variants for each position of the 11-mer peptide (460RLGFFGFDLQD470, the highlighted Gly represents the site of natural thioamidation). Additionally, each Phe was individually replaced with Tyr and Trp, while the position equivalent to Gly465 was also substituted with Val and Pro (SI Appendix, Table S2). These variants were assayed under two reaction conditions and then analyzed by endpoint MALDI-TOF-MS. These experiments established the positions equivalent to M. acetivorans McrA Asp467 and Gly465 as the most critical—variants at these locations nearly abolished thioamide formation even under forcing reaction conditions. The specificity for position Asp467 was further assessed by substitution with Gly, Glu, and Asn. The Gly variant was not processed by TfuA/YcaO, while the Glu/Asn variants were minimally processed. The positions equivalent to McrA Leu461, Phe463, Leu468, and Gln469 were additionally identified as important residues for recognition as their replacement severely impaired thioglycine formation (Fig. 3C and SI Appendix, Table S2). To define a minimal substrate for the M. acetivorans YcaO/TfuA proteins, we prepared truncations of the 11-mer peptide and, subsequently, evaluated the variants as substrates using MALDI-TOF-MS. Only one residue could be trimmed (equivalent to Asp470), as any further truncation abolished thioamidation (SI Appendix, Table S2).
Fig. 3.
Substrate scope analysis of the YcaOs. (A) Fluorescence polarization (FP) binding curve of M. acetivorans YcaO with the FITC-labeled M. acetivorans McrA 11-mer. Error bars represent SD of the mean (n = 3). (B) Summary of the binding constants obtained by competitive FP using Ala-substituted variants of the 11-mer using M. acetivorans YcaO. Errors on KD and Ki are the SEM calculated during regression analysis. (C) Summary of the substrate tolerance of the Ala-substituted variants of the 11-mer using reactivity-based analysis. Important recognition residues for M. acetivorans YcaO-TfuA, M. kandleri YcaO and M. jannaschii YcaO were determined from the results of MALDI-TOF-MS assays. Circles indicate the sites of Ala substitution in the M. acetivorans McrA 11-mer peptide (thioamidated Gly is shown in red), and the filled circles indicate the variants that inhibited processing in each case.
To bolster the results from our thioamidation assays, we probed the interaction between the McrA peptide fragments and M. acetivorans YcaO/TfuA. We labeled the N terminus of a synthetic McrA 11-mer (GG-RLGFFGFDLQD) with fluorescein isothiocyanate and measured the binding to M. acetivorans YcaO and TfuA using a fluorescence polarization (FP)-based method. We observed binding of the 11-mer peptide (KD ∼ 0.7 µM) to YcaO (Fig. 3A), but binding to TfuA was not detected. To elucidate the recognition residues on the substrate, we assessed the binding of all of the Ala-substituted variants of the 11-mer peptide toward YcaO by competitive FP. While the wild-type MBP-tagged 11-mer efficiently competed with the fluorescein-labeled 11-mer (Ki ∼ 0.7 µM), other variants were severely reduced in their ability to compete. For instance, Ala substitution of the position equivalent to Leu461 reduced affinity to the YcaO by 28-fold. Additional reductions were observed as follows: Gly462 (14-fold), Phe463 (>30-fold), Phe464 (11-fold), Gly465 (14-fold), Asp467 (12-fold), Leu468 (14-fold), and Gln469 (24-fold) (Fig. 3B). These binding data are well-aligned with the substrate selectivity results.

TfuA-Independent YcaOs Do Not Require a Partner Protein.

To investigate whether TfuA-independent YcaOs from methanogens are competent thioamidation catalysts, we heterologously expressed and purified the YcaOs from two hyperthermophilic methanogens (M. kandleri and M. jannaschii; SI Appendix, Fig. S4) in Escherichia coli. Upon assaying these YcaOs under conditions identical to that used for M. acetivorans, we observed ions consistent with thioamidation on the M. acetivorans 11-mer peptide, demonstrating that TfuA was not required for the reaction (SI Appendix, Fig. S10). Using the same panel of substrate variants of the 11-mer peptide from M. acetivorans, Phe464 and Gly465 were found to be essential residues for M. kandleri YcaO while Leu461, Phe463, and Phe466 contributed to a lesser extent (Fig. 3 and SI Appendix, Table S3). Replacement of Asp467 with Ala, Gly, Glu, or Asn did not adversely affect the activity. Moreover, Phe464 and Phe466 to Tyr variants were improved substrates for the M. kandleri YcaO. For M. jannaschii YcaO, important recognition residues were found to be Leu461, Phe463, Phe464, Gly465, Phe466, and Asp467 (SI Appendix, Table S4) while Phe464 and Phe466 to Tyr variants were better tolerated. Both YcaOs accepted shorter minimal substrates, relative to M. acetivorans YcaO, based on processing of the truncation variants, i.e., RLGFFGF for M. kandleri YcaO (SI Appendix, Table S3) and RLGFFGFD for M. jannaschii YcaO (SI Appendix, Table S4). Thioamidation activity with their native McrA 11-mer peptides (Fig. 4) was also confirmed by MS (SI Appendix, Figs. S11–S13).
Fig. 4.
MALDI-TOF-MS analysis of McrA 11-mer peptide thioamidation using M. jannaschii YcaO. The McrA fragment was from M. jannaschii with the residue naturally thioamidated (Gly447) shown in red. (A) Mass spectrum of the unmodified peptide, m/z 1,460 Da. (B) Mass spectrum of the peptide after reaction with the full array of reactants. (C) Identical to B but with the omission of YcaO. (D) Omission of ATP. (E) Omission of sodium sulfide (Na2S).

Mechanistic Analysis.

Azoline-forming YcaOs utilize ATP to activate the peptide backbone via formation of a phosphorylated hemiorthoamide (Fig. 1) (12, 24). Reaction monitoring of M. jannaschii YcaO thioamidation by 31P NMR detected ADP and phosphate (Pi) as the byproducts (SI Appendix, Fig. S14). This points toward a phosphorylation mechanism for the ATP-dependent amide bond activation. When the thioamidation reaction was carried out in [18O]-H2O, we primarily observed the formation of [16O4] Pi by 31P NMR, the identity of which was confirmed by spiking with authentic [16O4] Pi. Thus, bulk solvent is not responsible for ATP hydrolysis but rather the backbone amide oxygen. Together, these data support a direct activation role for ATP during thioamidation, as previously established for azoline-forming YcaOs (24). Kinetic parameters for the M. jannaschii YcaO were also determined using the native 11-mer peptide as the substrate via the purine nucleoside phosphorylase assay (24) with the following results: kcat = 0.89 ± 0.04 s−1; KM, peptide = 83 ± 8 µM; and kcat/KM = 10,722 M−1s−1. The ATP KM was 117 ± 11 µM (SI Appendix, Fig. S15).

Crystal Structures of TfuA-Independent YcaO.

Numerous attempts to crystallize the M. acetivorans YcaO/TfuA were unsuccessful owing to protein instability. To maximize the probability of obtaining an atomic-resolution structure of a thioamide-forming YcaO, we focused on hyperthermophilic TfuA-independent YcaO proteins. Heterologously expressed and purified M. kandleri and M. jannaschii YcaOs were monomeric and well-behaved in solution, as determined by size-exclusion chromatography (SI Appendix, Fig. S16). Crystals were obtained for both of these proteins in the presence of various nucleotides, but given that the M. kandleri YcaO structures diffracted to higher resolution (2.05–2.35 Å; SI Appendix, Table S5), we focused structural efforts on this protein. Phases were determined for M. kandleri YcaO using anomalous diffraction data collected from selenomethionine-labeled protein.
The structures of the M. kandleri YcaO (Fig. 5 and SI Appendix, Fig. S17) recapitulate the overall structure of known YcaOs as observed in LynD, an ATP-dependent peptide cyclodehydratase from cyanobactin biosynthesis (36) (PDB ID code: 4V1T) and the E. coli enzyme of unknown function (PDB ID code: 4Q86) (12). However, there are numerous insertions of secondary structural elements, as well as significant movement near regions of the nucleotide-binding pocket, which may have catalytic implications. Cocrystal structures with bound ADP-Mg2+ (2.05 Å resolution), ATP-Mg2+ (2.35 Å) and the nonhydrolyzable ATP analog, β,γ-methyleneadenosine triphosphate (ACP)-Mg2+ (2.3 Å), all show the nucleotide bound in a similar location. The constellation of residues that engage the nucleotide are similar to those observed in LynD and E. coli YcaO, but unlike these prior structures, M. kandleri YcaO only binds a single Mg2+. The side chain of Gln70 is within hydrogen bonding distance to N6 and N7 of the adenine, the O2′ and O3′ atoms of the ribose are within hydrogen bonding distance to the main chain N of Ala150 and the side chains of Arg12 and Glu81. The α-phosphate is engaged by Lys64 and Ser74 while the γ-phosphate is engaged by Arg168. Likewise, the Arg636/Arg286 of LynD/E. coli YcaO, which forms a salt bridge with the α-phosphate, is retained as Arg247 in M. kandleri YcaO. The catalytically requisite Mg2+ is engaged by Glu81 and Glu164, while Glu78, Gln160, and Glu167 are hydrogen bonded to solvent molecules that complete the Mg2+ coordination sphere. There are also notable differences in residues that engage the nucleotide, as Asn536/Asn187 that hydrogen bond to the N1 of the adenine in LynD/E. coli YcaO are replaced by Arg152, which is directed away from the active site of M. kandleri YcaO. The orientation of the nucleotide in the ATP-Mg2+ and ACP-Mg2+ structures positions the γ-phosphate toward a solvent-exposed region, which is mechanistically consistent with phosphorylation as opposed to adenylation (Fig. 6).
Fig. 5.
Structure of M. kandleri YcaO. (A) Overall structure of M. kandleri YcaO in complex with ACP-Mg2+ with secondary structural elements labeled as per SI Appendix, Fig. S17. Helices are colored in pink, and strands are shown in green. (B) Simulated annealing difference Fourier map (Fobs − Fcalc) contoured at 2.5σ showing the active site of M. kandleri YcaO-ADP-Mg2+. Active site residues implicated in catalysis are shown in stick (brown), along with the bound ADP (yellow) and Mg2+ (cyan sphere). (C) Simulated annealing difference Fourier map (Fobs − Fcalc) contoured at 2.5σ showing the active site of M. kandleri YcaO-ACP-Mg2+ complex.
Fig. 6.
Probable mechanistic routes for YcaO-dependent thioamidation. (A) Phosphorylation mechanism, in which the sulfide nucleophile attacks the target amide bond of the peptide first forming a tetrahedral intermediate. In the next step, the backbone oxyanion will attack the γ-phosphate of ATP to generate ADP and another thiolate phosphorylated intermediate, which will resolve by releasing inorganic phosphate (Pi) and the thioamidated peptide. (B) Adenylation mechanism in which the first tetrahedral intermediate will attack the α-phosphate of ATP to generate pyrophosphate (PPi) and the adenylated thiolate intermediate. This would then resolve by releasing AMP and the thioamidated peptide.
While the nucleotide-binding site is reminiscent of, but distinct from, that observed in LynD and E. coli YcaO (12), the remainder of the M. kandleri YcaO is entirely reorganized. Specifically, the thioamide-forming homologs lack several α-helices that are adjacent to the γ-phosphate in LynD, resulting in an open site where the peptide substrate would presumably bind. Additionally, helices α6 and α7 (SI Appendix, Fig. S17) are shifted away at the location of the γ-phosphate, resulting in a greater solvent-exposed active site. Lastly, a segment between helices α9 and α11 are disordered in all seven crystallographically independent copies of the polypeptide in every M. kandleri YcaO structure. The equivalent region in LynD abuts the γ-phosphate. As a result of the disorder, residues Ser265 through Arg268 of M. kandleri YcaO (SI Appendix, Fig. S18) are positioned directly in place of the terminal five residues of LynD that form the critical C-terminal PxPxP motif, which aids in organizing the active site residues of azoline-forming YcaO proteins (12, 25, 36).

Sequence and Structure-Guided Mutagenesis.

We validated the ATP-binding residues by conducting structure/function studies on the M. kandleri and M. jannaschii YcaOs based on sequence alignment (SI Appendix, Fig. S3) as well as the nucleotide-bound structure (Fig. 5). Ala replacements were prepared for all of the conserved polar residues for M. jannaschii YcaO and within 5 Å of the nucleotide pocket in M. kandleri YcaO. All variants were well-tolerated in terms of protein yield and stability (SI Appendix, Fig. S4). The effect on thioamide formation on the M. jannaschii McrA 11-mer peptide substrate (442RLGFYGYDLQD452) by the M. jannaschii YcaO variants were monitored by MALDI-TOF-MS (Table 1). Of the 16 Ala-substituted residues, two variants were as active as the wild type, five exhibited partial processing, while no activity was observed for the remaining nine variants (Table 1).
Table 1.
Summary of the M. jannaschii YcaO variants targeting the ATP-binding pocket that includes the substrate (M. jannaschii McrA 11-mer peptide) binding constants (KD) determined by the FP binding assay (SI Appendix, Fig. S19) and relative thioamide installation determined by endpoint MALDI-TOF-MS
M. jannaschii YcaO variantsKD, µMRelative processing
WT1.1 ± 0.08++
R11A1.4 ± 0.04++
D41A1.7 ± 0.06
K68A1.2 ± 0.05+
S78A3.3 ± 0.1+
E82A*9.3 ± 0.6
E85A*2.3 ± 0.06
E98A1.2 ± 0.03++
N155A8.9 ± 0.7
H169A1.1 ± 0.02+
E173A*1.1 ± 0.1
E176A*2.3 ± 0.9
R177A12.3 ± 0.5
D184A1.6 ± 0.04+
R161A1.0 ± 0.07+
E265A*1.1 ± 0.03
Q268A*1.2 ± 0.06
++, fully processed; +, partially processed; −, no product detected; ND, not determined due to unstable protein.
Residues within contact distance of the triphosphate group of ATP/Mg2+.
Supporting evidence for this mutational analysis comes from an FP assay (Table 1) in which perturbation in substrate binding was measured for all of the M. jannaschii YcaO variants (SI Appendix, Figs. S3 and S19). Upon Ala substitution of variants that coordinate ATP (E85A and E176A), minor perturbation in substrate binding was observed while a more pronounced effect was observed for E82A. Among the variants that showed partial processing, S78A exhibited a threefold reduction in KD. Two inactive variants, N155A and R177A, displayed severely reduced binding. Corresponding substitutions of conserved residues in M. kandleri YcaO were also prepared and their relative processing of the M. kandleri McrA 11-mer peptide substrate (443RLGFYGYALQD453) was evaluated by a MALDI-TOF-MS endpoint assay (SI Appendix, Fig. S3 and Table S6). Thioamidation activity was abolished when residues involved in ATP coordination (E78, E81, E164, and E167) and probable ATP/substrate coordination (E251 and Q254) were substituted with Ala (SI Appendix, Table S6). Reduced activity was observed for additional variants (K64A, S74A, and R247A) while activity was unaltered for the remainder of the positions examined. A similar trend was observed from an Ala replacement study on the conserved ATP-binding residues in M. acetivorans YcaO (E87A, E90A, and E201A) as well as additional conserved residues (R40A and D48A; SI Appendix, Fig. S3 and Table S7). In contrast to azoline-forming YcaO proteins, truncation of the C terminus of the M. acetivorans YcaO did not abolish the catalytic activity.


Thioamide-containing compounds are rare in biology. Compared with thioamide-containing nucleosides (32, 34), the biosynthesis of peptidic thioamides is poorly understood. Based on earlier studies on azoline-forming YcaOs, we predicted that thioamide formation on the peptidic backbone would also be an ATP-dependent, YcaO-catalyzed reaction (12). Our deletion studies in M. acetivorans provided experimental evidence for the involvement of the ycaO and tfuA genes in thioglycine formation in McrA (11). Herein, we report the successful reconstitution of peptidic thioamidation using McrA peptides, YcaO, and TfuA from M. acetivorans in the presence of ATP and Na2S as an external sulfide donor. Additional characterization will be necessary to determine if sulfur transferases are directly involved in MCR thioamidation in vivo, although there is literature precedence of direct use of inorganic sulfides for certain methanogen cellular functions (37, 38). Based on previous reports on YcaO partner proteins, we propose that TfuA may allosterically activate YcaO or assist in the delivery of sulfide to the substrate. To confirm or refute these proposals, additional experimentation will be required.
Using several variants of the M. acetivorans McrA 11-mer peptide substrate, followed by biochemical evaluation, we identified the residues that govern the site selectivity of thioamidation by M. acetivorans YcaO/TfuA. These data suggest a probable hydrogen bond from McrA-Asp467 to the YcaO that is abolished upon replacement with Ala or Gly. Truncation analysis identified Arg460-Gln469 as a minimal substrate under our reaction conditions, as no enzymatic processing was observed when the Gln469 codon was replaced with a stop codon.
To validate whether TfuA-independent YcaOs require any partner protein during thioamidation, we employed YcaOs from two hyperthermophilic methanogens (M. kandleri and M. jannaschii) in our assays. We demonstrated that these YcaOs, which are ∼40 residues shorter than the TfuA-dependent M. acetivorans YcaO, could install thioamide on peptidic substrates without requiring the assistance of any partner proteins. Using the same panel of M. acetivorans McrA 11-mer peptide (Arg460-Asp470) variants, we identified residues important for substrate recognition for both YcaOs. Interestingly, Asp467, a critical residue for processing by M. jannaschii YcaO, is replaced with Ala in the native M. kandleri McrA sequence (443RLGFYGYALQD453). Moreover, the variants of the 11-mer in which Phe464 and Phe466 were substituted with Tyr were more efficiently processed, which better reflects their native McrA sequences. Both YcaOs were capable of thioamide installation on shorter substrates compared with the case of M. acetivorans.
Akin to the azoline-forming YcaOs (12), thioamide-forming YcaOs employ a phosphorylation mechanism as opposed to an adenylation mechanism (Fig. 6). Using 31P NMR, we show that ADP and Pi are the reaction byproducts with the backbone amide oxygen being incorporated into the released Pi, as revealed by isotope-labeling experiments. Based on these studies, we put forward a mechanistic proposal for YcaO-catalyzed thioamide formation. Upon peptide substrate binding, an external source of sulfide will attack the target amide bond (here Gly465), generating a tetrahedral intermediate. The amide oxyanion will then attack the γ-phosphate of ATP, releasing ADP and forming a phosphorylated thiolate intermediate. This thermodynamically favorable step in which C-S bond formation is linked to ATP hydrolysis might be step wise or concerted. Next, the tetrahedral intermediate will collapse to form the thioamide and release Pi.
To date, structural data on nucleotide-bound YcaOs were previously limited to the cyanobactin LynD cyclodehydratase (36) and the orphan E. coli YcaO (25) for which a function has yet to be firmly established (12). The nucleotide-bound structures of the thioamide-forming M. kandleri YcaO display several features common to other YcaO proteins, but the structures also show notable differences. The nucleotide-binding site is largely retained and residues that interact with the nucleotide are equivalent to those observed in azoline-forming YcaOs (25). However, the remainder of the active site is largely divergent, resulting in a large, exposed area in the vicinity of the γ-phosphate. Moreover, several residues adjacent to this moiety are disordered and occlude the position of the C-terminal PxPxP motif critical for azoline-forming YcaOs (12). These observations suggest that either the thioamide-forming YcaOs contain a divergent substrate-binding site or that binding of the substrate results in a reorganization of the active site to one that is more catalytically competent. Based on activity assays with Ala replaced variants, we identified several important residues in the nucleotide-binding pocket for all of the three YcaOs that would be involved in ATP/substrate coordination during thioamidation.
Our data on enzyme-catalyzed thioamidation of ribosomal peptides illuminate a path forward for the elucidation and characterization of additional thioamide-containing natural product biosynthetic pathways as well as enable a deeper investigation into the native biological functions of this rare posttranslational modification.

Materials and Methods

Detailed methods are provided in SI Appendix, SI Materials and Methods. These cover experimental procedures for preparing and analyzing the proteins/peptides used, reaction conditions, and the analytical procedures. Protocols for the enzyme kinetics and crystallography data collection, structure determination, and refinement are also included.

Data Availability

Data deposition: The atomic coordinates have been deposited in the Protein Data Bank, www.wwpdb.org [PDB ID codes 6CIB (Mk-YcaO-ADP) and 6CI7 (Mk-YcaO-ACP)].


We thank William Metcalf for M. acetivorans genomic DNA, Keith Brister and staff at the Advanced Photon Source (Argonne National Laboratory) for facilitating data collection, and Graham Hudson and Tyler Maggio for their technical assistance. This work was supported in part by National Institutes of Health Grant GM097142 (to D.A.M.).

Supporting Information

Appendix (PDF)


S Scheller, M Goenrich, R Boecher, RK Thauer, B Jaun, The key nickel enzyme of methanogenesis catalyses the anaerobic oxidation of methane. Nature 465, 606–608 (2010).
U Ermler, W Grabarse, S Shima, M Goubeaud, RK Thauer, Crystal structure of methyl-coenzyme M reductase: The key enzyme of biological methane formation. Science 278, 1457–1462 (1997).
RK Thauer, A-K Kaster, H Seedorf, W Buckel, R Hedderich, Methanogenic archaea: Ecologically relevant differences in energy conservation. Nat Rev Microbiol 6, 579–591 (2008).
SJ Moore, et al., Elucidation of the biosynthesis of the methane catalyst coenzyme F430. Nature 543, 78–82 (2017).
K Zheng, PD Ngo, VL Owens, XP Yang, SO Mansoorabadi, The biosynthetic pathway of coenzyme F430 in methanogenic and methanotrophic archaea. Science 354, 339–342 (2016).
M Goubeaud, G Schreiner, RK Thauer, Purified methyl-coenzyme-M reductase is activated when the enzyme-bound coenzyme F430 is reduced to the nickel(I) oxidation state by titanium(III) citrate. Eur J Biochem 243, 110–114 (1997).
T Wongnate, et al., The radical mechanism of biological methane synthesis by methyl-coenzyme M reductase. Science 352, 953–958 (2016).
T Selmer, et al., The biosynthesis of methylated amino acids in the active site region of methyl-coenzyme M reductase. J Biol Chem 275, 3755–3760 (2000).
J Kahnt, et al., Post-translational modifications in the active site region of methyl-coenzyme M reductase from methanogenic and methanotrophic archaea. FEBS J 274, 4913–4921 (2007).
T Wagner, J Kahnt, U Ermler, S Shima, Didehydroaspartate modification in methyl-coenzyme M reductase catalyzing methane formation. Angew Chem Int Ed Engl 55, 10630–10633 (2016).
DD Nayak, N Mahanta, DA Mitchell, WW Metcalf, Post-translational thioamidation of methyl-coenzyme M reductase, a key enzyme in methanogenic and methanotrophic Archaea. eLife 6, e29218 (2017).
BJ Burkhart, CJ Schwalen, G Mann, JH Naismith, DA Mitchell, YcaO-dependent posttranslational amide activation: Biosynthesis, structure, and function. Chem Rev 117, 5389–5456 (2017).
KL Dunbar, DH Scharf, A Litomska, C Hertweck, Enzymatic carbon-sulfur bond formation in natural product biosynthesis. Chem Rev 117, 5521–5577 (2017).
PG Arnison, et al., Ribosomally synthesized and post-translationally modified peptide natural products: Overview and recommendations for a universal nomenclature. Nat Prod Rep 30, 108–160 (2013).
Y Hayakawa, K Sasaki, K Nagai, K Shin-ya, K Furihata, Structure of thioviridamide, a novel apoptosis inducer from Streptomyces olivoviridis. J Antibiot (Tokyo) 59, 6–10 (2006).
L Frattaruolo, R Lacret, AR Cappello, AW Truman, A genomics-based approach identifies a thioviridamide-like compound with selective anticancer activity. ACS Chem Biol 12, 2815–2822 (2017).
L Kjaerulff, et al., Thioholgamides: Thioamide-containing cytotoxic RiPP natural products. ACS Chem Biol 12, 2837–2841 (2017).
GE Kenney, AC Rosenzweig, Chemistry and biology of the copper chelator methanobactin. ACS Chem Biol 7, 260–268 (2012).
GE Kenney, AC Rosenzweig, Genome mining for methanobactins. BMC Biol 11, 17 (2013).
OD Hensens, G Albers-Schönberg, Total structure of the highly modified peptide antibiotic components of thiopeptin. J Antibiot (Tokyo) 36, 814–831 (1983).
MS Puar, et al., Sch 18640. A new thiostrepton-type antibiotic. J Am Chem Soc 103, 5231–5233 (1981).
T Lincke, S Behnken, K Ishida, M Roth, C Hertweck, Closthioamide: An unprecedented polythioamide antibiotic from the strictly anaerobic bacterium Clostridium cellulolyticum. Angew Chem Int Ed Engl 49, 2011–2013 (2010).
M Izawa, T Kawasaki, Y Hayakawa, Cloning and heterologous expression of the thioviridamide biosynthesis gene cluster from Streptomyces olivoviridis. Appl Environ Microbiol 79, 7110–7113 (2013).
KL Dunbar, JO Melby, DA Mitchell, YcaO domains use ATP to activate amide backbones during peptide cyclodehydrations. Nat Chem Biol 8, 569–575 (2012).
KL Dunbar, et al., Discovery of a new ATP-binding motif involved in peptidic azoline biosynthesis. Nat Chem Biol 10, 823–829 (2014).
BA Schulman, JW Harper, Ubiquitin-like protein activation by E1 enzymes: The apex for downstream signalling pathways. Nat Rev Mol Cell Biol 10, 319–331 (2009).
KL Dunbar, JI Tietz, CL Cox, BJ Burkhart, DA Mitchell, Identification of an auxiliary leader peptide-binding protein required for azoline formation in ribosomal natural products. J Am Chem Soc 137, 7672–7677 (2015).
BJ Burkhart, GA Hudson, KL Dunbar, DA Mitchell, A prevalent peptide-binding domain guides ribosomal natural product biosynthesis. Nat Chem Biol 11, 564–570 (2015).
CJ Schwalen, et al., In-vitro biosynthetic studies of bottromycin expand the enzymatic capabilities of the YcaO superfamily. J Am Chem Soc 139, 18154–18157 (2017).
L Franz, S Adam, J Santos-Aberturas, AW Truman, J Koehnke, Macroamidine formation in bottromycins is catalyzed by a divergent YcaO enzyme. J Am Chem Soc 139, 18158–18161 (2017).
CL Cox, JR Doroghazi, DA Mitchell, The genomic landscape of ribosomal peptides containing thiazole and oxazole heterocycles. BMC Genomics 16, 778 (2015).
CM Wright, GD Christman, AM Snellinger, MV Johnston, EG Mueller, Direct evidence for enzyme persulfide and disulfide intermediates during 4-thiouridine biosynthesis. Chem Commun (Camb), pp. 3104–3106 (2006).
S Coyne, et al., Control of plant defense mechanisms and fire blight pathogenesis through the regulation of 6-thioguanine biosynthesis in Erwinia amylovora. ChemBioChem 15, 373–376 (2014).
S Coyne, et al., Biosynthesis of the antimetabolite 6-thioguanine in Erwinia amylovora plays a key role in fire blight pathogenesis. Angew Chem Int Ed Engl 52, 10564–10568 (2013).
C Chen, H Huang, CH Wu, Protein bioinformatics databases and resources. Methods Mol Biol 1558, 3–39 (2017).
J Koehnke, et al., Structural analysis of leader peptide binding enables leader-free cyanobactin processing. Nat Chem Biol 11, 558–563 (2015).
Y Liu, M Sieprawska-Lupa, WB Whitman, RH White, Cysteine is not the sulfur source for iron-sulfur cluster and methionine biosynthesis in the methanogenic archaeon Methanococcus maripaludis. J Biol Chem 285, 31923–31929 (2010).
Y Liu, LL Beer, WB Whitman, Methanogens: A window into ancient sulfur metabolism. Trends Microbiol 20, 251–258 (2012).

Information & Authors


Published in

Go to Proceedings of the National Academy of Sciences
Go to Proceedings of the National Academy of Sciences
Proceedings of the National Academy of Sciences
Vol. 115 | No. 12
March 20, 2018
PubMed: 29507203


Data Availability

Data deposition: The atomic coordinates have been deposited in the Protein Data Bank, www.wwpdb.org [PDB ID codes 6CIB (Mk-YcaO-ADP) and 6CI7 (Mk-YcaO-ACP)].

Submission history

Published online: March 5, 2018
Published in issue: March 20, 2018


  1. thioamide
  2. methanogen
  3. posttranslational modification
  4. YcaO
  5. methyl-coenzyme M reductase


We thank William Metcalf for M. acetivorans genomic DNA, Keith Brister and staff at the Advanced Photon Source (Argonne National Laboratory) for facilitating data collection, and Graham Hudson and Tyler Maggio for their technical assistance. This work was supported in part by National Institutes of Health Grant GM097142 (to D.A.M.).


This article is a PNAS Direct Submission.



Nilkamal Mahanta
Department of Chemistry, University of Illinois, Urbana, IL 61801;
Carl R. Woese Institute for Genomic Biology, University of Illinois, Urbana, IL 61801;
Andi Liu
Carl R. Woese Institute for Genomic Biology, University of Illinois, Urbana, IL 61801;
Department of Microbiology, University of Illinois, Urbana, IL 61801;
Department of Biochemistry, University of Illinois, Urbana, IL 61801;
Satish K. Nair
Carl R. Woese Institute for Genomic Biology, University of Illinois, Urbana, IL 61801;
Department of Biochemistry, University of Illinois, Urbana, IL 61801;
Center for Biophysics and Quantitative Biology, University of Illinois, Urbana, IL 61801
Department of Chemistry, University of Illinois, Urbana, IL 61801;
Carl R. Woese Institute for Genomic Biology, University of Illinois, Urbana, IL 61801;
Department of Microbiology, University of Illinois, Urbana, IL 61801;


To whom correspondence should be addressed. Email: [email protected].
Author contributions: N.M., A.L., S.D., S.K.N., and D.A.M. designed research; N.M., A.L., and S.D. performed research; N.M., A.L., and S.D. contributed new reagents/analytic tools; N.M., A.L., S.D., S.K.N., and D.A.M. analyzed data; and N.M., A.L., S.K.N., and D.A.M. wrote the paper.

Competing Interests

The authors declare no conflict of interest.

Metrics & Citations


Note: The article usage is presented with a three- to four-day delay and will update daily once available. Due to ths delay, usage data will not appear immediately following publication. Citation information is sourced from Crossref Cited-by service.

Citation statements



If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. Simply select your manager software from the list below and click Download.

Cited by


    View Options

    View options

    PDF format

    Download this article as a PDF file


    Get Access

    Login options

    Check if you have access through your login credentials or your institution to get full access on this article.

    Personal login Institutional Login

    Recommend to a librarian

    Recommend PNAS to a Librarian

    Purchase options

    Purchase this article to get full access to it.

    Single Article Purchase

    Enzymatic reconstitution of ribosomal peptide backbone thioamidation
    Proceedings of the National Academy of Sciences
    • Vol. 115
    • No. 12
    • pp. 2843-E2901







    Share article link

    Share on social media

    Further reading in this issue