Engineered reactivity of a bacterial E1-like enzyme enables ATP-driven modification of protein and peptide C termini

Abstract

In biological systems, ATP provides an energetic driving force for peptide bond formation, but protein chemists lack tools that emulate this strategy. Here, we develop an ATP-driven platform for C-terminal activation and peptide ligation based on MccB, a bacterial ancestor of ubiquitin-activating (E1) enzymes. We show that MccB can act on non-native substrates to generate an O-AMPylated electrophile that reacts with exogenous nucleophiles to form diverse C-terminal functional groups including thioesters, a versatile class of biological intermediates that have been exploited for protein C-terminal bioconjugation. By mining the natural diversity of the MccB family, we identify both epitope-specific and more promiscuous MccBs. We show that epitope-specific MccB activity can be directed toward specific proteins of interest to enable high-yield, ATP-driven protein bioconjugation, while promiscuous MccB activity can be deployed for synthesis of peptide thioester substrates for bioconjugation. Our method mimics the chemical logic of biological peptide bond synthesis for high-yield in vitro manipulation of protein structure with molecular precision.

Introduction

In living systems, ATP provides an energetic driving force for peptide bond synthesis. For example, activation of the α-carboxylate of amino acids or protein/peptide C termini by adenylation enables formation of ester and thioester intermediates that function in diverse pathways including ribosomal protein synthesis¹, non-ribosomal peptide synthesis², and the ubiquitination cascade³. Cleavage of ATP to AMP and pyrophosphate (PPi) provides a large thermodynamic driving force (ΔG°′ = −45.6 kJ/mol for hydrolysis) for otherwise unfavorable biosynthetic reactions⁴. In both protein translation and non-ribosomal peptide synthesis, amino acids are initially activated by ATP-dependent adenylation to hydrolytically unstable aminoacyl-AMPs, which are converted to aminoacyl-tRNAs or aminoacyl-peptidyl carrier proteins that can be used in protein synthesis or non-ribosomal peptide synthesis, respectively. In contrast, enzyme-generated adenylates have not been used as activated intermediates for in vitro protein bioconjugation by bioorganic chemists^5,6. Instead, the most widely adopted protein bioconjugation methods rely on reversible enzyme-catalyzed transpeptidation reactions^7–13 or non-enzymatic pre-synthesis of activated precursors^14–16 to provide a driving force for peptide bond formation¹⁷.

The E1-like (ThiF) enzyme superfamily is comprised of structurally and mechanistically related proteins that share a common mechanistic step involving adenylation of a C-terminal α-carboxylate to generate a reactive peptidyl-O-AMP mixed anhydride electrophile^18,19. Based on its capacity to react with diverse nucleophiles, this shared intermediate provides biological systems with access to protein and peptide C-terminal thioesters³, thiocarboxylates^20,21, succinimides^22,23 and (iso)peptide bonds^24,25 that function in biological processes ranging from ubiquitination to biosynthesis of metabolites including thiamin, molybdopterin, and ribosomally synthesized and post-translationally modified peptide (RiPP) natural products. Despite their versatility, E1-like enzymes have not been developed for in vitro applications in peptide bond synthesis as many of the best studied E1-like enzymes require ubiquitin-like proteins as their substrates. In contrast to many other E1-like enzymes, E. coli MccB, which natively functions in formation of a C-terminal phosphoramidate linkage in biosynthesis of the RiPP natural product microcin C7, accepts a short heptapeptide as its substrate^22,26. This feature makes MccB a promising candidate tool for protein bioconjugation. However, MccB’s native reaction chemistry does not couple C-terminal adenylation to a peptide bond formation reaction that would be useful for protein bioconjugation.

Here, we report the mechanism-guided design of an MccB-based toolbox for C-terminal activation and protein modification. We show that E. coli MccB has a latent capacity to catalyze C-terminal adenylation on non-native substrates. The resultant C-terminal peptidyl-O-AMP electrophile can react with a variety of exogenous nucleophiles, including hydrazines, alkoxyamines, amines, and thiols, to form diverse C-terminal functional groups. We develop the Thioesterification C-terminal Handle (TeCH-tag), a sequence that enables MccB-catalyzed, ATP-driven synthesis of protein C-terminal thioesters, an important class of protein bioconjugation intermediates that were previously only directly accessible via a method that relies on engineered self-splicing inteins. By mining the natural diversity of the MccB family, we identify two additional MccB/TeCH-tag pairs that are mutually orthogonal to each other and to the E. coli system, as well as a more promiscuous MccB homolog. We apply the MccB/TeCH-tag system for high-yield, ATP-dependent protein bioconjugation via expressed protein ligation and enzyme-catalyzed expressed protein ligation. We find that MccB/TeCH-tag is compatible with other bioconjugation enzymes such as sortase, enabling synthesis of dual N- and C-terminally functionalized protein conjugates. We also develop a more promiscuous homolog of MccB for synthesis of Ub-derived peptide thioester substrates that can be used as substrates for the Lysine Acylation using Conjugating Enzymes (LACE) system. These strategies mimic the chemical logic of peptide bond synthesis that is widespread in biology for high-yield in vitro synthesis of protein bioconjugates that will advance our understanding of biological systems.

Results

Mechanism-guided design of a C-terminal modification tool

E. coli MccB is a member of the E1-like enzyme superfamily that catalyzes conversion of the C-terminal Asn residue of its heptapeptide substrate, MccA (MRTGNAN), to an isoasparagine (isoAsn)-AMP phosphoramidate (Fig. 1A, B)^22,26. As its first mechanistic step, MccB is proposed to catalyze C-terminal adenylation (or O-AMPylation) of MccA, producing an MccA-O-AMP intermediate that is captured by the β-carboxamido nitrogen group of the C-terminal Asn (N7) residue to form a succinimide intermediate (Fig. 1C). This succinimide is subsequently N-AMPylated and hydrolyzed to form the C-terminal isoAsn-AMP product²². This modified heptapeptide is further tailored by phosphoramidate aminopropylation^27,28 and acts as a ‘Trojan horse’ antibiotic that is cleaved by an endogenous protease in target cells to form isoAsn-AMP, an aspartyl-adenylate mimic that inhibits the aspartyl-tRNA synthetase^29,30. While most families within the E1-like superfamily act on ubiquitin-like β-grasp fold proteins with a C-terminal Gly residue^18,19, MccB family enzymes are unique in that they accept as their substrates short, genetically encoded peptides terminating with C-terminal Asn (Fig. 1A, B)^31,32. We hypothesized that if we substituted the C-terminal Asn of the MccA substrate with another amino acid, MccB would retain the ability to O-AMPylate the non-native substrate and that in the absence of a cis nucleophile in the substrate, the peptidyl-O-AMP electrophile could react with an exogenous nucleophile for C-terminal functionalization.

Figure 1. MccB catalyzes C-terminal O-AMPylation of non-native substrates.

(A) MccB enzymes are members of the E1-like/ThiF superfamily, which have in common the formation of a C-terminal peptidyl-O-AMP intermediate. (B) Members of the E1-like superfamily that act on C-termini generally recognize ubiquitin-like proteins that terminate in Gly as their substrates; the MccB family is an exception as it acts on short peptides (MccAs) that terminate in Asn. (C) The reaction catalyzed by MccB in its native context. (D) The C-terminal residue of E. coli MccA was varied to the 19 non-native amino acids and MccB’s ability to catalyze C-terminal O-AMPylation was assayed using a colorimetric assay for pyrophosphate. (E) MccA-N7A, N7G, N7S, and N7T stimulate MccB-catalyzed pyrophosphate formation (left column), while only wt MccA supports formation of a stably AMPylated product detectable by HPLC-MS (right column). (F) Steady-state kinetic parameters for MccB-catalyzed pyrophosphate release when wt MccA or MccA variants are used as substrates. Kinetic parameters are given as value ± standard error.

To test the hypothesis that MccB retains a latent capacity to O-AMPylate peptide substrates lacking the conserved C-terminal Asn residue, we synthesized a small library of non-native MccA peptides in which the C-terminal residue was varied to the 19 amino acids other than Asn (MccA-N7X). Because the peptidyl-O-AMP product of MccB-catalyzed O-AMPylation is expected to be hydrolytically unstable, we initially screened the ability of these MccA variants to stimulate ATP consumption by MccB using an enzyme-coupled assay to detect formation of the pyrophosphate (PP_i) by-product (Fig. 1D)^33,34. We observed substantial stimulation of PP_i formation when MccA-N7A, N7G, N7S, or N7T were used as substrates for MccB (Fig. 1E, Fig. S1). We determined the steady-state kinetic parameters for MccB-catalyzed pyrophosphate formation with each of these substrates and found that the catalytic efficiencies (k_cat/K_M) were similar to or exceeded the k_cat/K_M measured for wild-type MccA (Fig. 1F, Fig. S2). We next tested whether this ATP consumption resulted in the formation of any stably AMPylated product using an HPLC-MS assay. While we observed a clear +329.0485 Da peak corresponding to AMPylation when wild-type MccA was used as a substrate, no stably AMPylated product was observed for the MccA-N7A, N7G, N7S, or N7T variants (Fig. 2A, B, Fig. S3, Table S1). These results are consistent with a previous study that showed that N7 is required for microcin C7’s antimicrobial activity³⁵. Together these data raise two possibilities: 1) Binding of MccA-N7X variants to MccB stimulates unproductive ATP hydrolysis; or 2) MccA-N7X variants bind to MccB and are O-AMPylated, but the resulting mixed anhydride is hydrolytically unstable and cannot be detected by HPLC-MS.

Figure 2. MccB catalyzes formation of a peptidyl-O-AMP intermediate that can react with exogenous nucleophiles.

(A) MccA-O-AMP undergoes reaction with its C-terminal Asn side chain to form a succinimide intermediate and subsequently a stably N-AMPylated product that can be detect by LC-MS (top). In contrast, MccA-N7G-O-AMP is expected to undergo hydrolysis on a timescale incompatible with LC-MS detection. (B) LC-MS analysis of MccB reactions with wild-type MccA (left) or MccA-N7G (right) as a substrate. (C) Hypothesized MccA-N7G-O-AMP reactivity with exogenous nucleophiles. (D) MccB-catalyzed hydrazine modification of MccA-N7G. ESI-MS spectra of unmodified MccA-N7G (left) and MccA-N7G (250 μM) incubated with MccB (5 μM), ATP (5 mM), and hydrazine (150 mM) reveal that MccA-N7G-O-AMP can react with exogenous nucleophiles. (E) Radial heatmaps show the percent conversion of MccA to nucleophile-modified MccA for a panel of alkoxyamine, hydrazine, amine, and thiol nucleophiles. Percent conversion was calculated using peak areas from reactant and product extracted ion chromatograms.

We hypothesized that if an electrophilically activated O-AMPylated C terminus is formed, an exogenously added nucleophile, hydrazine, could compete with water to attack the electrophilic group, resulting in MccA-N7X bearing a C-terminal hydrazide that could be distinguished from the unmodified peptide based on its mass (+14.0269 Da) (Fig. 2C). Such experiments have been used previously to detect electrophilic enzymatic intermediates including anhydrides and thioesters^36,37. To test this hypothesis, we included 150 mM hydrazine in reactions containing MccB, ATP, and MccA-N7G (M+H⁺ = 706.3306). HPLC-MS analysis of these reactions revealed formation of a new peak corresponding to formation of a C-terminal hydrazide (M+H⁺ = 720.3583) that was dependent on MccB, ATP, and hydrazine (Fig. 2D). Similar results were obtained using hydroxylamine as the trapping nucleophile (Fig. 2D, Fig. S4). These data support the hypothesis that MccB catalyzes O-AMPylation of non-Asn C-terminal residues and provide direct evidence for the O-AMPylated intermediate in the proposed mechanism of MccB.

We next sought to define features of the nucleophile that impact the efficiency of capture of the O-AMPylated intermediate. We initially screened a panel of hydroxylamines, hydrazines, and amines by measuring their ability to modify the C terminus of MccA-N7G in an MccB-dependent manner using HPLC-MS (Fig. 2E, Fig. S4, Table S2). Among hydroxylamine nucleophiles, increasing the size of the O-substituent decreased the efficiency of C-terminal modification. Similarly, we found that a recently reported phenylhydrazine-based nucleophilic probe (N-(but-3-yn-1-yl)-4-(2-hydrazineylethyl)benzamide)³⁸ was unable to modify the C terminus of MccA-N7G. These results suggest that nucleophile size is an important factor in the efficiency of MccB-catalyzed MccA-N7G modification and raise the hypothesis that the intermediate is not released from the enzyme but reacts with exogenous nucleophiles within the enzyme active site. In contrast to the efficient capture observed with hydroxylamine and hydrazine, small primary and secondary amines modified MccA inefficiently (0–26%) (Fig. 2E, Fig. S4). These results suggest that nucleophile strength is also an important factor in efficient capture of the O-AMPylated intermediate.

We next tested whether MccB could catalyze ATP-dependent C-terminal thioesterification of MccA-N7G in the presence of thiol nucleophiles. We tested a panel of thiols comprised of N-acetyl-l-cysteine, N-aceylcysteamine, DTT, sodium 2-mercaptothane sulfonate (Mesna), and 4-mercaptophenylacetic acid (MPAA). With the exception of MPAA, we found that all of these thiol nucleophiles modified MccA-N7G in 76–98% yield (Fig. 2E, Fig. S4). We attribute the inability of MPAA to modify MccA-N7G to the constraints on nucleophile size and structure that are imposed by the enzyme active site. These results demonstrate that the MccA-N7G-O-AMP intermediate can be efficiently transferred to thiol nucleophiles for C-terminal thioesterification.

Protein C-terminal thioesterification and ligation with MccB

Protein and peptide C-terminal thioesters are key reactive species in both biological and chemical peptide bond formation pathways. For example, in the canonical ubiquitination cascade, an E1 catalyzes C-terminal O-AMPylation of ubiquitin (Ub), activating Ub for nucleophilic attack by a Cys side chain to generate an E1-bound Ub C-terminal thioester (denoted E1~Ub). E1~Ub then undergoes transthioesterification with a Cys side chain of an E2 ubiquitin-conjugating enzyme to form E2~Ub. Finally, an E3 ubiquitin ligase binds both E2~Ub and the substrate to catalyze S-to-N acyl transfer to a Lys side chain from the substrate to form an isopeptide bond³⁹. Similarly, in native chemical ligation (NCL), a peptide C-terminal thioester undergoes transthioesterification with a peptide bearing an N-terminal Cys residue, follow by S-to-N acyl shift to form a native peptide bond. While the initial transthioesterification step in NCL is reversible, the S-to-N acyl shift that forms the amide bond is irreversible, driving the reaction to high yield.

We found that C-terminal thioesters synthesized by MccB could undergo both transthioesterification with MPAA, a thiol that cannot efficiently capture the MccA-N7G-O-AMP electrophile directly (Extended Data Fig. 1, Fig. S5, Fig. S6). This result suggests that transthioesterification occurs outside the enzyme active site and is not subject to the same steric restraints as the initial thioesterification reaction. We also found that when L-cysteine is used as the thiol donor, the thioester underwent S-to-N acyl shift to form a peptide bond (Extended Data Fig. 1, Fig. S7). Encouraged by these results, we sought to test whether N-terminal Cys peptides could similarly capture MccA-N7G-O-AMP and undergo S-to-N acyl shift to enable a peptide ligation reaction. We first tested a Cys-Trp dipeptide nucleophile and observed no modification of MccA-N7G, suggesting that the larger nucleophile is unable to access MccA-N7G-O-AMP within the MccB active site (Fig. S8). However, when we included both Mesna and Cys-Trp in the reaction, MccA-N7G was converted to MccA-N7G-CW in 82% yield in 4 hours (Extended Data Fig. 1, Fig. S8). This result suggests that MccA-N7G can undergo MccB-catalyzed thioesterification with Mesna, followed by transthioesterification and S-to-N acyl shift with Cys-Trp in one pot, enabling enzyme-catalyzed NCL starting from an unactivated peptide with a C-terminal carboxylate.

C-terminal protein thioesters are key intermediates in expressed protein ligation (EPL), a powerful protein semisynthesis method that involves NCL between a recombinantly expressed protein C-terminal thioester and a synthetic N-terminal Cys peptide to form a native peptide bond⁴⁰. Prior applications of EPL have relied on fusion of the protein to be modified with an engineered intein to generate the protein C-terminal thioester⁶. Encouraged by the ability of MccB to catalyze MccA-N7G thioesterification to activate it for C-terminal peptide ligation, we hypothesized that MccA-N7G could be developed as a tag for C-terminal protein thioesterification. Previous functional characterization of MccB demonstrated that an N-terminal maltose binding protein (MBP) fusion of MccA is recognized as a substrate by MccB and can be modified with an N-P bond to AMP³². To test whether MccA-N7G could similarly enable protein C-terminal thioesterification, we fused GFP to MccA-N7G (a sequence that we term the Thioesterification C-terminal Handle, or TeCH-tag) (Fig. 3A, Fig. S9). In the presence of 5 mM Mesna and 5 mM ATP, we found that MccB catalyzed near-quantitative conversion of GFP-TeCH-tag to the Mesna thioester with 30 min (Fig. 3B, Fig. S10, Fig. S11). In comparison, a GFP-TeCH-Mxe GyrA intein fusion protein yielded a mixture of GFP-TeCH-Mes (40%) and hydrolyzed GFP-TeCH (60%) (Fig. S12). When the partially hydrolyzed intein-generated thioester was treated with MccB, ATP, and Mesna, it was converted nearly quantitatively to the Mes thioester (Fig. S13). The MccB/TeCH-tag system therefore provides an efficient method to generate C-terminal protein thioesters.

Figure 3. Fusion of the Thioesterification C-terminal Handle (TeCH-tag) to proteins enables MccB-catalyzed, ATP-dependent formation of C-terminal thioesters.

(A) Fusion of the TeCH-tag to GFP for C-terminal thioesterification. (B) MccB catalyzes ATP-dependent thioesterification of GFP-TeCH-tag within 30 min. (C) MccB catalyzes C-terminal thioesterification of TeCH-tag fusions of MBP, the catalytic domain of protein tyrosine phosphatase 1B (PTP1B_1–321), protein L, an α-GFP recombinant antibody, and an EGFR-targeting affibody. The * indicates an α-gluconylated form of protein L that is an artifact of His-tag purification. (D) MccB catalyzes C-terminal thioesterification of GFP-TeCH-tag with cysteine, with subsequent S-to-N acyl shift leading to formation of a peptide bond as evidenced by the maleimide reactivity of the bioconjugate. (E) GFP-TeCH-tag can be modified by expressed protein ligation via a Mesna thioester intermediate in a one-pot reaction with MccB, ATP, Mesna, and the peptide CGAGS-azidoalanine.

To examine the general utility of the MccB/TeCH-tag system, we appended the TeCH-tag to a diverse panel of proteins, including MBP; the catalytic domain (residues 1–321) of protein-tyrosine phosphatase 1B (PTP1B_1–321); an anti-GFP recombinant antibody (α-GFP rAb); protein L; and an endothelial growth factor receptor (EGFR)-targeting affibody (zEGFR). We treated these TeCH-tag fusion proteins (50 μM) with MccB (5 μM), ATP (5 mM), and Mesna (5 mM) and found that they were thioesterified in near-quantitative yield within 1–16 hours (Fig. 3C, Fig. S14–S23). While GFP, MBP, and PTP1B_1–321 were completely converted to thioester in 1 h, zEGFR required 6 h, and protein L and α-GFP rAb required 16 h. Notably, the α-GFP rAb, based on a scaffold derived from the therapeutic antibody Trastuzumab⁴¹, contains five disulfide bonds linking its light and heavy chains. We found that the rAb could be thioesterified in quantitative yield under conditions that omit reducing agents other than Mesna, keeping the disulfide bonds required for antibody function intact. These results demonstrate that the MccB/TeCH-tag system is broadly applicable for generating protein C-terminal thioesters.

We next examined whether the MccB/TeCH-tag system could be deployed to expand the toolbox for C-terminal thioester formation in the context of EPL. We initially tested whether MccB could catalyze incorporation of single Cys residue at the C terminus of a Cys-free variant of GFP (cfGFP)⁴² via an amide bond. We incubated cfGFP (50 μM) with MccB (5 μM), ATP (5 mM), and Cys (5 mM) and found that Cys was ligated to cfGFP in in 99% yield in 1 h (Fig. 3D, Fig. S9, S24, S25). Following Cys incorporation, the Cys side chain was quantitatively modified with biotin-maleimide or Cy5-maleimide, demonstrating that S-to-N acyl shift had occurred to produce a free Cys side chain (Fig. 3D, Fig. S26, S27). We tested whether a workflow involving thioesterification with Mesna, transthioesterification with an N-terminal Cys peptide, and S-to-N acyl shift could be applied for ligation of cfGFP to a synthetic N-terminal Cys peptide bearing an azide functional group. We performed a reaction in which TeCH-tagged cfGFP was incubated with MccB, ATP, Mesna, and peptide with the sequence CGAGS-3-azido-l-Ala (Table S3, Fig. S28). We found that cfGFP-TeCH-tag could be quantitatively modified with the azide-bearing peptide (Fig. 3E, Fig. S29). The azide modified protein could then undergo strain-promoted azide-alkyne cycloaddition (SPAAC) with dibenzocyclooctyne (DBCO)-biotin (Fig. 3E, Fig. S30). MccB can thus serve as a catalyst for protein C-terminal thioesterification to enable EPL.

Natural MccA/MccB diversity encompasses orthogonal pairs

To examine the sequence specificity of MccB for the MccA substrate, we synthesized a library of peptides in which each of the amino acid in each position of MccA was varied to the 19 non-native canonical amino acids (Fig. 4A, Fig. S31, Table S4). We then measured formation of the C-terminally N-AMPylated phosphoramidate product for each of the MccA variants to query the stringency of MccB’s sequence specificity. We found that only wild-type MccA was fully converted to product over 16 h. Consistent with our previous results, no product formation was observed if the seventh position was varied to an amino acid other than Asn. Outside the C-terminal residue, substitutions in the first two positions of MccA had the largest effect on MccB activity, with significant product formation observed only for the M1W, M1Y, R2K, and R2L variants. At positions 3–6, significant product formation was observed for 6–8 different substitutions in each position. However, none of these variants were converted to product in quantitative yield despite the long reaction time, suggesting that MccB is an epitope-specific enzyme. Steady-state kinetics analysis of Ala variant peptides revealed that MccB’s lower activity on MccAs with substitutions in positions 1–3 is mainly attributable to decreases in k_cat (Fig. S32). Our data are broadly consistent with a previous study that examined the effect of substituting MccA positions 2–7 on microcin C7 production and antimicrobial activity in vivo and found that positions 4–6 are most tolerant of substitutions³².

Figure 4. Natural MccA/MccB diversity encompasses orthogonal enzymes for C-terminal protein modification.

(A) A positional scanning peptide library revealed that MccB is an epitope-specific enzyme. (B) The MccB enzyme family harbors numerous homologs that act on substrates distinct from E. coli MccA. Enzymes from H. pylori, L. johnsonii,and H. somni are highlighted in cyan. The E. coli, H. pylori, and L. johnsonii MccB were found to be mutually orthogonal enzymes. (C) The E. coli, H. pylori, and L. johnsonii enzymes catalyze C-terminal thioesterification of their respective MccA-N7G sequences and are orthogonal to one another. (D) LC-MS analysis of GFP-EcTeCH, GFP-HpTeCH, and GFP-LjTeCH shows that MccA-N7Gs from E. coli, H. pylori, and L. johnsonii can be deployed as TeCH-tags for C-terminal protein modification. (E) MccBs from E. coli, H. pylori, and L. johnsonii modify only proteins tagged with their cognate TeCH-tags. (F) In a mixture of GFPs fused to TeCH-tags from E. coli, H. pylori, and L. johnsonii, the MccB homologs from E. coli, H. pylori, and L. johnsonii selectively modify their cognate TeCH-tags.

Epitope-specific bioconjugation enzymes enable modification of proteins to probe their functions, to discover inhibitors and drugs, to immobilize them for catalysis, and conjugate them to cytotoxic drugs, among many other applications^8,10,43. However, their application can be limited by the relatively small number of available modification epitopes that restricts their use in the synthesis of more complex bioconjugates and in orthogonally targeting multiple different proteins in a mixture^10,43,44. As a result, there is strong interest in identifying or engineering orthogonal enzyme/substrate pairs for protein modification. We hypothesized that the natural diversity of MccA and MccB might encompass mutually orthogonal enzyme-substrate pairs (Fig. 4B). Bioinformatic analyses have revealed that many bacterial genomes that encode MccB homologs also encode MccA-like peptides in the same gene cluster^31,32. These analyses annotated 31 distinct, previously unknown heptapeptide MccA sequences as well as 14 longer putative MccAs. Based on the stringent sequence specificity of E. coli MccB, we hypothesized that the MccB homologs that recognize these distinct substrate sequences might be orthogonal to E. coli MccB and to one another. To test this hypothesis, we measured the activity of MccBs from E. coli (EcMccB), Helicobacter pylori (HpMccB), Lactobacillus johnsonii (LjMccB), and Histophilus somni (HsMccB) toward both their native and non-cognate substrates using an enzyme-coupled assay to measure PPi release (Fig. 4B, Fig. S33). All four enzymes had the highest level of activity on their native substrates. While EcMccB, HpMccB, and LjMccB had <1% activity on non-cognate sequences, HsMccB exhibited a detectable amount of activity (18–27% of HsMccA) on all three non-cognate substrates tested. These results suggested that EcMccB/EcMccA, HpMccB/HpMccA, and LjMccB/LjMccA might be useful as mutually orthogonal enzyme-substrate pairs.

We next sought to test whether HpMccB and LjMccB could catalyze C-terminal thioesterification of HpMccA-N7G and LjMccA-N7G, respectively. We incubated each enzyme with its cognate MccA-N7G, ATP, and Mesna. Using LC-MS analysis, we found that each enzyme converted its cognate substrate to the C-terminal thioester in >97% yield within 16 h (Fig. 4C, Fig. S34, S35). We next tested whether EcMccB, HpMccB, and LjMccB catalyzed C-terminal thioesterification of MccA-N7Gs from other species. We did not detect thioesterification of non-cognate substrates by the EcMccB or LjMccB over 16 h, while a small of amount of thioesterification (13%) of LjMccA catalyzed by HpMccB was observed (Fig. 4C, Fig. S35). These results suggest that these three enzyme-substrate pairs possess a high degree of mutual orthogonality in terms of their ability to catalyze C-terminal thioesterification.

To test whether HpMccA-N7G and LjMccA-N7G could be used as TeCH-tags for C-terminal protein thioesterification, we fused these sequences to the C terminus of GFP. We found that both enzymes converted their cognate TeCH-tagged GFPs to C-terminal thioesters in the presence of 50 mM Mesna (Fig. 4D, Fig. S36–S39). To test the orthogonality of the three MccB/TeCH-tag systems that we developed, we examined the ability of each MccB to modify GFP-EcTeCH, GFP-HpTeCH, and GFP-LjTeCH. We found that each MccB was only able to modify its cognate GFP-TeCH (Fig. 4E, Fig. S36–S39). Next, we tested whether each MccB could selectively modify its cognate GFP-TeCH in a mixture of GFP-EcTeCH, GFP-HpTeCH, and GFP-LjTeCH (Fig. 4F, Fig. S40–S43). We observed that each of the MccBs only modified its cognate substrate, even at high (50 mM) Mesna concentrations. EcMccB/EcTeCH, HpMccB/HpTeCH, and LjMccB/LjTeCH therefore represent three orthogonal C-terminal modification enzymes that greatly expand the available toolbox of epitope specific bioconjugation enzymes.

MccB-generated thioesters enhance the bioconjugation toolbox

Although EPL is a powerful and widely adopted tool for protein semisynthesis⁶, enzymatic strategies for C-terminal bioconjugation offer alternative approaches with different advantages depending on the specific application and target protein. For example, compared to EPL, the enzyme-catalyzed EPL approach increases sequence flexibility at the ligation junction⁴⁵, while sortagging simplifies the bioconjugation strategy for applications in which sequence scars are tolerated⁸. Based on the utility of these approaches, we sought to evaluate how the application of MccB-generated C-terminal protein thioesters could enhance enzyme-catalyzed C-terminal bioconjugation strategies.

We first sought to apply MccB in the context of enzyme-catalyzed EPL for epitope-specific bioconjugation. In the enzyme-catalyzed EPL approach, a C-terminal protein thioester is used as a substrate for the engineered peptide ligase subtiligase^14,46, which has broad N-terminal specificity and eliminates the requirement for Cys at the ligation site. While this method has been applied to study phosphoregulation of the tyrosine phosphatase PTEN by introduction of phosphoresidues at specific sites^45,47, yields were limited by a competing subtiligase-catalyzed thioester hydrolysis reaction. We hypothesized that application of MccB for ATP-dependent thioester generation in this context would drive yields higher because it would enable thioester regeneration from the inactivated hydrolysis product (Fig. 5A).

Figure 5. MccB enables ATP-dependent thioester formation and regeneration for high-yield enzyme-catalyzed expressed protein ligation.

(A) Enzyme-catalyzed expressed protein ligation is limited by subtiligase-catalyzed hydrolysis of the thioester substrate, generating a dead-end product. We used MccB for ATP-dependent thioester formation and regeneration, enabling reactivation of the dead-end hydrolytic product. (B) High yield one-pot MccB- and subtiligasecatalyzed ATP-dependent peptide ligation to GFP-TeCH. (C) MccB- and subtiligase-catalyzed ATP-dependent peptide ligation to TeCH-tag fusions of MBP, the catalytic domain of protein tyrosine phosphatase 1B (PTP1B_1–321), protein L, an α-GFP recombinant antibody, and an EGFR-targeting affibody. The * indicates an α-gluconylated form of protein L that is an artifact of His-tag purification. (D) MccB- and subtiligase-catalyzed peptide ligation and strain promoted azide-alkyne cycloaddition were used to synthesize α-GFP rAb-Cy3 for staining of a HEK293T cell line engineered for doxycycline-inducible expression of cell surface GFP. (E) Dual N- and C-terminal labeling of TeCH-tagged MBP with 5-FAM-LPETGG and AFAGAGS-azidoAla using eSrtA and MBP/subtiligase under one-pot (center) or telescoping (right) conditions.

To test this hypothesis, we incubated TeCH-tagged GFP (50 μM) with MccB (5 μM), ATP (5 mM), Mesna (5 mM), subtiligase (5 μM), and AFAGAGS-azidolysine (5 mM, Table S3), which contains an azide for downstream modification using click chemistry. We found that GFP could be modified efficiently with this peptide (Fig. 5B, Fig. S44) A timecourse indicated that t_1/2 for the reaction was 0.9±0.1 h (Fig. S45). GFP variants bearing TeCH-tag sequences derived from H. pylori and L. johnsonii could also be efficiently modified by enzyme-catalyzed EPL (Extended Data Fig. 2, Fig. S46–S48). To test whether the efficiency of the ligation reaction depends on the protein substrate, we tested the reaction on our panel of TeCH-tag fusion proteins (MBP, PTP1B_1–321, α-GFP rAb, protein L, and zEGFR) (Fig. 5C, Fig. S49–S53). We found that all the proteins tested could be efficiently ligated to AFAGAGS-azidolysine in high yields, highlighting the general utility of MccB for driving subtiligase-catalyzed peptide ligation.

To test the utility of bioconjugates synthesized with MccB/TeCH-tag enzyme-catalyzed EPL method in a biological context, we constructed a HEK293T cell line that expresses cell surface GFP under the control of a tetracycline/doxycycline-inducible promoter. We synthesized a Cy3-modified α-GFP rAb by using MccB and subtiligase to ligate AFAGAGS-azidolysine onto the C-terminal TeCH-tag fused to the heavy chain, followed by SPAAC with DBCO-Cy3 (Fig. S54). We then stained doxycycline (Dox)-induced cells and uninduced cells with the α-GFP rAb-Cy3 conjugate. We observed robust Cy3 staining that colocalized with GFP in the Dox-induced cells, while neither GFP signal nor Cy3 signal was observed in uninduced cells (Fig. 5D). These results demonstrate the utility of MccB- and subtiligase-catalyzed bioconjugation for incorporating probes into antibodies while maintaining their ability to bind their targets.

Similar to sortagging, the TeCH-tag/MccB/subtiligase system achieves protein bioconjugation through formation of a peptide bond. We therefore sought to compare the efficiency of MccB/subtiligase-catalyzed protein bioconjugation to sortase-catalyzed bioconjugation using the engineered sortase variant eSrtA⁴⁸ . We initially replaced the TeCH-tag in our GFP construct with the eSrtA recognition sequence LPETGG. In contrast to the near-quantitative conversion to the desired ligation product catalyzed by MccB, eSrtA (2.5 μM) catalyzed 75% ligation of GFP-LPETGG (50 μM) to a triglycine nucleophile (GGG, 5 mM), while 25% of GFP-LPETGG was cyclized based on the presence of an N-terminal Gly residue that could serve as an eSrtA substrate (Extended Data Fig. 3, Fig. S55–S57). This intramolecular reaction could not be suppressed even when the GGG concentration was increased to 10 mM (Fig. S58). We hypothesize that cyclization was effectively suppressed in the MccB/subtiligase reaction in the presence of 5 mM ligation partner because the ligation product is no longer a substrate for MccB. In contrast, the eSrtA-catalyzed reaction is reversible, and formation of the desired ligation product is governed by the position of the equilibrium between the intermolecular and intramolecular products. Although this intramolecular reaction could be blocked by using an alternative N-terminal sequence (Extended Data Fig. 3, Fig. S59–61), the ability of MccB/subtiligase-catalyzed bioconjugation to avoid this reaction highlights the utility of using irreversible steps to drive the reaction along an intended trajectory.

We next assessed whether MccB/subtiligase and eSrtA can function in combination for dual functionalization of a single protein. As a test substrate, we used MBP modified at the N terminus with Gly-Ser and at the C terminus with a TeCH tag. We incubated this protein (25 μM) with eSrtA (2.5 μM), MccB (5 μM), subtiligase (5 μM), AFAGAGS-azidoAla (5 mM, MccB/subtiligase substrate), and 5-FAM-LPETGG (2 mM, eSrtA substrate) for 12 h at room temperature. In the presence of all three enzymes, we observed 72% conversion to dual modified protein, while only 5-FAM-LPET modification was observed in the absence of MccB, only AFAGAGS-azidoAla modification was observed in the absence of eSrtA, and 5-FAM-LPET/Mes modification was observed in the absence of subtiligase (Extended Data Fig. 4, Fig. S62–S65). To optimize conversion to the dual N- and C-terminally modified product, we next tried a telescoping approach in which GS-MBP-TeCH (25 μM) was incubated with MccB (5 μM), subtiligase (5 μM), and AFAGAGS-azidoAla (5 mM, MccB/subtiligase substrate) for 4 h at room temperature, followed by addition of eSrtA (2.5 μM) and 5-FAM-LPETGG (2 mM) and incubation for an additional 4 h at room temperature (Extended Data Fig. 4, Fig. S66–S70). Under telescoping conditions, we observed near-quantitative conversion of GS-MBP-TeCH to a species modified with both the FAM- and azide-bearing peptides. Notably, dual N- and C-terminal labeling using orthogonal sortase variants is currently limited by the incomplete orthogonality of these engineered variants in terms of the nucleophilic ligation partners that they accept. The application of MccB/subtiligase therefore complements and expands the existing toolkit for complex bioconjugation applications.

MccB-catalyzed Ub-derived peptide thioester synthesis

In our screen for mutually orthogonal MccA/MccB pairs, we observed that H. somni MccB exhibited more cross-reactivity than homologs from E. coli, H. pylori, and L. johnsonii (Fig. 4B). We wondered whether HsMccB’s expanded substrate tolerance might make it suitable for synthesis of thioesters derived from peptide sequences divergent from the native HsMccA sequence. We chose to examine whether HsMccB could be deployed to synthesize Ub-derived peptide thioesters for lysine acylation using conjugating enzymes (LACE), a recently developed bioconjugation strategy⁴⁹ . In LACE, a lysine within a genetically encoded tag of 4–13 residues (the LACE tag) is recognized by the E2 SUMO-conjugating enzyme Ubc9 and modified with a peptide thioester derived from Ub (Fig. 6A). The typical peptide thioester sequence motif recognized by Ubc9 is comprised of the six C-terminal residues of Ub, LRLRGG, with the final three amino acids representing a minimal motif for Ub loading.

Figure 6. Enzymatic synthesis of ubiquitin-derived peptide thioesters using an MccB homolog with relaxed substrate specificity.

(A) Synthetic ubiquitin-derived peptide C-terminal thioesters are substrates for Ubc9 in the lysine acylation with conjugating enzymes (LACE) strategy. (B) MccB-catalyzed synthesis of peptide C-terminal thioesters. (C) MccB from Histophilus somni (cyan, HsMccB) acts on a basic substrate and exhibited lower substrate specificity in our screen of MccB homologs. (D) HsMccB efficiently converts Ubc9 substrate peptides for LACE to Mes thioesters (left), while AcCysNHMe (right) was a less effective thiol donor. (E) Application of MccB-generated thioesters for acylation of an internal lysine side chain in LACE-tagged GFP. Left, acylation using 300 μM LRLRGG-Mes thioester. Right, acylation using 750 μM MccB-generated MLGLRGG-Mes thioester.

We hypothesized that HsMccB might be useful for thioesterification of these Ub-derived peptide based on its promiscuity as well as the basic nature of both the HsMccA-N7G and LRLRGG peptides (Fig. 6B,C). We synthesized a panel of peptides that introduce one amino acid variation at a time to convert HsMccA-N7G to a Ubc9 substrate and tested whether they could be thioesterified by HsMccB. We found that HsMccB catalyzed efficient (>99%) thioesterification of HsMccA-N7G (MRGRRLG), HsLACE1 (MRGRRGG), HsLACE2 (MRGLRGG), and HsLACE3 (MLGLRGG) with Mesna as a thiol donor (Fig. 6D). HsLACE4 (MLRLRGG, <1% thioesterification) and LRLRGG (18% thioesterification) were poor substrates. We also tested AcCysNHMe, the most widely used thiol donor for synthesis of Ubc9 peptide thioester substrates and found that it was a poor thiol donor for HsMccB-catalyzed thioesterification (Fig. 6D).

We next tested the ability of our panel of HsMccB-generated Mes thioesters to serve as substrates for Ubc9-catalyzed lysine acylation of a LACE tag introduced into an internal site of GFP (following D173). In our initial screen, we found that efficient modification of the LACE tag only occurred when the peptide thioester contained at least the four C-terminal residues of Ub (LRGG) (Extended Data Fig. 5). We therefore proceeded with MLGLRGG, initially attempting one-pot and telescoped HsMccB/Ubc9 GFP-LACE modification reactions, but we observed low conversion (1–15%) to modified GFP-LACE (Extended Data Fig. 5). We attribute these low yields to inhibition of the Ubc9-catalyzed reaction by the excess free thiol required for HsMccB thioester synthesis (Extended Data Fig. 5). We next incorporated C18 spin column cleanup of the crude reaction to remove excess thiol into our workflow. After thiol removal, peptide thioesters were used directly in the Ubc9 GFP-LACE modification reaction and gave 94% modification of the LACE-tagged protein, similar to a synthetic peptide thioester control, LRLRGG-Mes (Fig. 6E, Fig. S73–76). HsMccB therefore provides an alternative enzymatic route to peptide thioesters that are key reagents for chemical biology and that have previously been accessible only through traditional chemical synthesis. This enzymatic approach lowers the barrier for deployment of bioorganic chemistry approaches such as LACE tag for interdisciplinary scientists who may lack extensive expertise in synthetic chemistry.

Discussion

We designed the MccB/TeCH-tag system to mimic the chemical logic of peptide bond synthesis in biological systems to drive protein and peptide bioconjugation reactions to high yield. MccB/TeCH-tag can be used in the context of enzyme-catalyzed EPL for ATP-dependent thioester regeneration, driving the reaction equilibrium toward the desired ligation product and away from the dead-end thioester hydrolysis product formed by adventitious subtiligase reactivity. Our system avoids the reversibility of transpeptidases, such as sortase and asparaginyl endopeptidases, in which the desired ligation products are also transpeptidase substrates. Because this limitation is based on the position of the equilibrium, it cannot be overcome through transpeptidase engineering. Although depsipeptide (ester) and thiodepsipeptide (thioester) substrates have been used to drive transpeptidation, these substrates must be chemically synthesized, cannot be regenerated, and are mainly useful for N-terminal rather than C-terminal labeling.

The MccB/TeCH-tag system couples ATP cleavage to C-terminal activation via formation of a peptidyl-O-AMP for high-yield in vitro protein modification. Although acyl-O-AMPs are reactive electrophiles analogous to the acid chlorides and acid anhydrides often used in organic synthesis, they have not typically been viewed as modular intermediates that can be deployed for synthesis of modified proteins and peptides. We took advantage of our understanding of the enzymatic reaction mechanism of MccB to design a system for peptidyl-O-AMP synthesis that can be integrated with protein chemistry toolbox in modular fashion for bioconjugate synthesis. Introduction of the N7G substitution to MccA abolishes its ability to serve as a precursor to the antimicrobial compound microcin C7, but still supports the formation of a C-terminally O-AMPylated electrophile. We showed that MccA-N7G-O-AMP can react with alkoxyamines, hydrazines, amines, and thiols to form oximes, hydrazides, amides, and thioesters, respectively. We anticipate that this reactivity can easily be extended to other nucleophiles for synthesis of C-terminally modified proteins. For example, ammonia could be used as a nucleophile for modification of proteins by C-terminal amidation, which has recently been shown to target proteins for ubiquitin modification by SCF/FBXO31 and proteasomal degradation⁵⁰. Other classes of nucleophiles such as alcohols could also capture the MccA-N7G-O-AMP intermediate for installation of protein C-terminal esters, which are present in prenylated proteins including Ras GTPases⁵¹.

We show that capture of MccA-N7G-O-AMP with thiol nucleophiles is particularly useful in protein bioconjugation as it converts the hydrolytically unstable peptidyl-O-AMP to a kinetically stable yet thermodynamically activated thioester. In biology, C-terminal thioesters function in enzymatic catalysis^52,53, serve as intermediates in protein splicing^54,55, and enable the installation of post-translational modifications including ubiquitin and ubiquitin-like proteins^3,19,56. Although nature has evolved several strategies to generate C-terminal thioesters, only one, intein-mediated protein splicing, had previously been harnessed as a tool for producing recombinant protein C-terminal thioesters that serve as versatile intermediates for synthesis of chemically tailored proteins⁶. Recombinant C-terminal thioesters can be deployed for native chemical ligation to N-terminal Cys peptides in expressed protein ligation (EPL)⁴⁰ or can be used as substrates for the engineered peptide ligase subtiligase in enzyme-catalyzed EPL⁴⁵. These strategies have enabled precise manipulation of protein structure to advance our understanding of a broad range of biological questions⁶. The MccB/TeCH-tag system expands the toolbox for direct C-terminal thioester synthesis from unactivated protein α-carboxylates and therefore represents a broadly applicable tool for protein bioconjugation.

Our results indicate that MccBs from E. coli, H. pylori, and L. johnsonii are epitope-specific, while MccB from H. somni is more promiscuous. Epitope-specific bioconjugation enzymes are valuable tools for protein modification, enabling installation of probes, payloads, and modifications that cannot be genetically encoded. In previous work, bacterial sortases have been widely applied for epitope-specific bioconjugation based on their selectivity for an LPXTG motif¹². However, synthesis of complex bioconjugates can be limited by the relatively small number of available epitope-enzyme pairs, and few orthogonal sortase/sorting motif pairs have been identified in nature⁵⁷. As a result, development of new orthogonal sortases has required intensive protein engineering efforts^43,44. In contrast, the natural diversity of MccA/MccB pairs enabled us to readily generate two orthogonal tools for epitope-specific protein bioconjugation. We also developed the more promiscuous H. somni for enzymatic synthesis of ubiquitin-derived peptide thioesters for the LACE system from unactivated peptides. Although we characterized four MccB/TeCH-tags, there are at least 31 distinct annotated heptapeptide MccA sequences with potential utility for protein bioconjugation^31,32. Future genome mining and experimental approaches to identify orphan MccB substrates have the potential to further expand the number of available MccB/TeCH-tag systems to enable the design of tailor-made bioconjugation enzymes that recognize user-defined sequences as well as more promiscuous enzymes for general C-terminal thioesterification.

Advances in our ability to construct modified proteins have expanded the frontiers of our understanding of how post-translational modifications regulate transcription⁵⁸, transduce cellular signals^59,60, and go awry in neurodegenerative disease^61–63; and have propelled our ability to probe biochemical and biophysical function through the installation of chemical probes and payloads that cannot be genetically encoded^8,64,65. The MccB/TeCH-tag system provides a method for C-terminal protein and peptide activation that vastly expands the toolkit for protein bioconjugation. MccB/TeCH-tag is broadly useful based on its ability to drive peptide bond-forming reactions to high yield by coupling them to ATP cleavage and can be integrated with existing protein chemistry technologies to fuel biological discovery.

Methods

Key chemicals and materials.

Reagents screened for nucleophilic capture of MccB-activated C termini are listed in Table S2. Maleimide, Dibenzocyclooctyne-PEG4-biotin, and AlaPhe dipeptide were purchased from Sigma Aldrich. TCEP hydrochloride was purchased from Gold Biotechnology. Cyanine5 maleimide was purchased from Lumiprobe. EZ-Link-maleimide-PEG2-biotin was purchased from Thermo Fisher Scientific. Protected amino acids, 1-hydroxybenzotriazole, 4-alkoxybenzyl alcohol resins, and Rink amide resin for solid-phase peptide synthesis were purchased from Chem Impex International.

Solid phase peptide synthesis.

MccA C-terminal variant peptides with C-terminal carboxylate groups were synthesized using fluorenylmethyloxycarbonyl (Fmoc) chemistry on 4-alkoxy-benzyl-alcohol resin preloaded with the required C-terminal amino acid (Chem Impex International). All other peptides with C-terminal amides were synthesized using Rink Amide resin (Chem Impex International). Fmoc amino acids with reactive side chains were protected with acid-labile protecting groups as follows: Asp (OtBu); Glu (OtBu); His(Trt); Lys(Boc); Asn(Trt); Gln(Trt); Arg(Pbf); Ser(tBu); Thr(tBu); Trp(Boc); Tyr(tBu). Fmoc groups were deprotected via 30-minute incubation in 20% methylpiperidine in DMF (20% v/v). Coupling steps were performed with 5 molar equivalents of the appropriate Fmoc-protected amino acid, 5 molar equivalents of diisopropylcarboiimide (DIC), and 5 molar equivalents of 1-hydroxy-benzotriazole (HOBt). Completed peptides were cleaved from the resin via incubation in a cocktail containing 95% trifluoroacetic acid, 2.5% triisopropylsilane, and 2.5% water. Peptides were concentrated under a stream of nitrogen and precipitated with 10 volumes of diethyl ether. Precipitated peptides were washed with additional diethyl ether and allowed to dry. The resulting crude product was purified using an Agilent 1260 Infinity II HPLC fitted with a semi-preparative ZORBAX Eclipse XDB-C18, 9.4 × 250 mm, 5 μm column. Crude products were separated using a 30-minute gradient from 0 to 100% B (A=0.1% trifluoroacetic acid in water, B=acetonitrile). Selected fractions were lyophilized, resuspended in water, and stored at −20°C. The positional scanning peptide library for MccB specificity characterization in Fig. 4 was purchased from Peptide2.0. ESI-MS data for synthetic peptides are shown in the Supplementary Information (Fig. S28, S46) or raw data have been deposited to Dryad under DOI: 10.5061/dryad.c59zw3rkb (MccA positional scanning peptide library)⁶⁶.

Mass spectrometry.

Electrospray ionization liquid chromatography mass spectrometry (ESI-LC-MS) analysis was performed on an Agilent 6230B time of flight (TOF) mass spectrometer. Samples containing peptide substrates were separated on an Agilent ZORBAX Eclipse XDB-C18, Solvent Saver Plus, 3 × 150 mm, 3.5 μm column using a 5-minute gradient from 0 to 100% B (A=0.1% formic acid in water, B=acetonitrile). Extracted ion chromatograms were generated using Agilent MassHunter Qualitative Analysis v10.0 and Agilent TOF Quantitative Analysis v11.0. Samples containing intact protein substrates were separated on a PLRP-S 1000 Å, 50 ×1 mm, 5 μm column at 80°C using a 3.9-minute gradient from 20 to 60% B (A=0.1% formic acid in water, B=acetonitrile). The maximum entropy charge deconvolution algorithm in Agilent MassHunter BioConfirm v10.0 was used to determine the neutral mass of intact proteins. For mixtures of TeCH-tagged proteins, the pMod algorithm was applied for charge deconvolution in Agilent MassHunter BioConfirm v10.0

Molecular biology and plasmid construction.

Plasmids were constructed using standard Gibson assembly cloning methods with E. coli XL10 as the cloning host. Oligonucleotides were purchased from Integrated DNA Technologies, and plasmid sequences were confirmed via Sanger Sequencing performed by Quintara Biosciences or Functional Biosciences. Plasmid maps have been deposited in Dryad under DOI: 10.5061/dryad.c59zw3rkb⁶⁶.

pBH4-His-TEV-MccB.

E. coli codon-optimized genes encoding MccBs from E. coli, H. pylori, L. johnsonii, and H. somni were purchased from Integrated DNA Technologies. The genes were inserted into the pBH4 vector between the BamHI and NotI restriction sites using Gibson assembly to generate constructs with an N-terminal His tag followed by a TEV protease cleavage site.

pBH4-His-TEV-cysteine-free (cf)-GFP-MccA fusions.

Cysteine-free eGFP was amplified from ss-cfSGFP2 (Addgene #37535). Primers (pBH4-His-Tev-cfGFP F1 and pBH4-His-Tev-cfGFP-EcMccA-N7G R1 or pBH4-His-Tev-cfGFP-LjMccA-N7G R1 or pBH4-His-Tev-cfGFP-HpMccA-N7G R1, Table S5) were used to add a C-terminal linker and MccA or MccA-N7G sequence. PCR products were inserted into pBH4 between BamHI and NotI restriction sites using Gibson assembly to generate a construct with an N-terminal His tag and a TEV protease cleavage site.

pBH4-His-TEV-MBP-EcTeCH.

MBP was amplified from pRK793 (Addgene #8827, a gift from Dr. David Waugh) using primers that added a C-terminal linker and TeCH tag (GGGSMRTGNAG) (pBH4-His-Tev-MBP F1 and pBH4-His-Tev-MBP-EcMccA-N7G R1, Table S5). The PCR product was inserted between BamHI and NotI restriction sites using Gibson assembly to generate a construct with an N-terminal His tag and a TEV protease cleavage site.

pBH4-His₆-TEV-PTP1B_1–321.

An E. coli codon-optimized gene encoding the PTP1B catalytic domain (residues 1–321) followed by a TeCH tag was purchased from Twist Bioscience. The gene was inserted between the BamHI and NotI sites of pBH4.

pET28a-His₆-protein L-EcTeCH.

An E. coli codon-optimized gene encoding His6-protein L with a C-terminal TeCH tag was purchased from Twist Bioscience. The gene was inserted between the NcoI and NdeI sites of pET28a.

pPSL937-anti-GFP-rAB-HC-TeCH.

pPSL937-anti-GFP⁶⁷ rAb was a gift from James A. Wells. A gene encoding the final 19 amino acids of the heavy chain of the rAb followed by a TeCH tag was purchased from Twist Bioscience. The gene was inserted between the two SalI sites of pPSL937.

pBH4-His-TEV-zEGFR-EcTeCH.

A gene encoding zEGFR followed by a C-terminal TeCH tag was purchased from Integrated DNA Technologies. The gene was inserted between the BamHI and NotI sites to generate a construct with an N-terminal His tag followed by a TEV protease cleavage site.

pBS42-pre-pro-stabiligase-His₆.

Construction of pBS42-pre-pro-stabiligase-His₆ was described previously. The plasmid was a gift from James A. Wells.

pET29-eSrtA.

pET29a-eSrtA (Addgene #75144) was a gift from David Liu.

pcDNA5/FRT/TO-Igκ-eGFP-PDGFRTM.

A synthetic gene encoding eGFP with an N-terminal Igκ signal peptide and a C-terminal PDGF receptor β-chain transmembrane domain to target eGFP to the cell surface was ordered from Twist Bioscience. The gene was inserted between the NcoI and NotI sites of pcDNA5/FRT/TO (Thermo Fisher) using Gibson assembly.

pTXB1-cfGFP-TeCH-Mxe.

pTXB1 was obtained as part of the IMPACT kit from New England Biolabs (catalog no. E6901S). cfGFP-TeCH was amplified from pBH4-cfGFP-EcMccA-N7G and inserted into the NdeI and SapI sites of pTXB1.

pBH4-His-TEV-Ubc9.

An E. coli codon-optimized gene encoding Ubc9 was purchased from Integrated DNA Technologies. The gene was inserted between the BamHI and NotI sites of pBH4 using Gibson assembly to generate a construct with an N-terminal His tag followed by a TEV protease cleavage site.

pBH4-His-TEV-GFP-D173-LACE tag.

Genes encoding previously reported GFP constructs with a minimal or full-length LACE tag sequence following D173 were purchased from Integrated DNA Technologies. The genes were inserted between the NcoI and NotI sites of pBH4.

Protein expression and purification.

Protein expression and purification methods for each protein used in this study are described below. Following purification, the purity of proteins was analyzed by SDS-PAGE and ESI-MS (Fig. S9, S12, S14, S16, S18, S20, S36, S38, S55, S59, S74, S77–S84).

E. coli MccB

MGHHHHHHDYDIPTTENLYFQGSMDYILGRYVKIARYGSGGLVGGGGKEQYVEDLALWENIIKTAYCFITPSSYTAALETVNIPEKDFSNCFRFLKENFFIIPSEYNNSTENNRYSRNFLHYQSYGANPVLVQDKLKDAKVVILGCGGIGNHVSVILATSGIGEIILIDNDQIENTNLTRQVLFSENDVGKNKTEVIKRELLKRNSEISVSEIALNINDYTDLHKVPEADIWVVSADHPFNLINWVNKYCVRANQPYINAGYVNDIAVFGPLYVPGKTGCYECQKVVADLYGSEKENIDHKIKLINSRFKPATFAPVNNVAAALCAADVIKFIGKYSEPLSLNKRIGIWSDEIKIHSQNMGRSPVCSVCGNRM

ESI-MS data for the purified protein are shown in Fig. S78.

Chemically competent E. coli BL21(DE3) cells were transformed with pBH4-His-TEV-E. coli MccB for overexpression. LB starter cultures (15 mL, 50 μg/mL carbenicillin) were grown overnight and used to inoculate 1 L LB cultures (50 μg/mL carbenicillin). Cultures were incubated at 37°C with vigorous shaking until OD₆₀₀ reached ~0.6. Cultures were chilled on ice for 15 minutes prior to adding 0.1 mM IPTG (isopropyl-β-D-thiogalactopyranoside). Cultures were shaken an additional 24 hours at 16°C. Cell pellets were harvested by centrifugation at 4°C, resuspended in 40 mL lysis buffer (25 mM Tris, 500 mM NaCl, 10 mM MgCl2, pH 8.0). Cells were lysed by three passes through an Emulsiflex microfluidizer at 15000 psi, and the resulting lysate was centrifuged at 8000 × g for 15 minutes at 4°C. Ni-NTA resin (1 mL) was added to the clarified lysate and His-tagged proteins were allowed to bind for 1 hour at 4°C with gentle rocking. The Ni-NTA resin was collected by centrifugation at 500 × g for 5 min, transferred to a 15 mL column, and washed with 15 mL wash buffer (20 mM Tris pH 8.0, 500 mM NaCl, 25 mM imidazole). Protein was eluted with 5 mL elution buffer (20 mM Tris pH 8.0, 500 mM NaCl, 200 mM imidazole) and dialyzed against wash buffer overnight at 4°C. The dialyzed protein was then concentrated using Amicon centrifugal filter units (10,000 MWCO) and further purified by size-exclusion chromatography (Superdex Hiload 75, GE Healthcare) using storage buffer (25 mM Tris pH 8.0, 50 mM NaCl, 1 mM DTT, 10% glycerol). For MccB/subtiligase experiments, glycerol was omitted from size-exclusion chromatography and storage buffers. Collected fractions were analyzed by Coomassie-stained SDS-PAGE. Fractions containing MccB were pooled, concentrated, and flash frozen in liquid nitrogen. Protein concentrations were determined using absorbance at 280 nm and the protein extinction coefficient as calculated using ProtParam.

H. pylori MccB

MGHHHHHHDYDIPTTENLYFQGSMQWYQTSFSACVGQTDTENIIGLGTYQYCVDHNEFEKSLKLLVFLRMKKRMAEIKSFMETSKIEHNIFDKLVANKLITSFILNPNDEQNFKNHLFIDLMSSKPELTIDNFKRTIFIIIGCGGIGNFVSYALASFYPKKLILLDKDTVDFSNLNRQFLFDKNYISQYKTSAIKQALSSRFSINIETVDDFASEDNLEEIFSKHKKENLFGIVSGDNPNTVQLATRFFCKCRIPFLNIGYLNDISLIGPFYIPSLSCCPFCHNSFALDDKKDGDENLDICLI

ESI-MS data for the purified protein are shown in Fig. S79.

L. johnsonii MccB

MGHHHHHHDYDIPTTENLYFQGSMFYKTSYLATGGCSNHQGILGVGTKQYFVSEADYLKSLKILDFLLNKKTYDEVIKFCEKNNINKSIFDTLVEHNLIVKENLYVEKKDDLNFKNKLYFHALGLNGNALAKEFADTTFVIVGCGGIGNFISFAIGSLSPRKIELIDGDKIEKSNLNRQFLFTENDIGKYKVDVLKKNLVERNNKLSISEYKEYVSKEVLHNIFEQNKKNKTLVILSGDSFSALSLTAKACVKSEIPFLNIGYLNDISAIGPFYIPGISSCPFCHNALSISDDISSGHNESKILEDRINANNEAPSSFTNNALAASMGIADIIEFLSHNYERINSLNKRFGINSATFEKYVLEVNRDRKCEICSHGE

ESI-MS data for the purified protein are shown in Fig. S80.

H. somni MccB

MGHHHHHHDYDIPTTENLYFQGSMKYITSKHVFFDYLNENEFVIGIGSNQEITNNKDYFNNCLNLCYFCINPKSISEILSFIKDNNIDILYFDKMKKMKFITKEIIDFNDRYSRNHLYYNALGYKIYDIQNKISKSHILIVGAGGIGNICSYLLGTIGIKKLSIIDDDIVEESNLNRQFLFREKDINKNKVETIKRELLSIRKDIIIDIFPEKLNKSILDKISQIDLVICSADDEYCIDMINEFCCFNKIPLINVGYLNDISVIGPFYIPKLEYSCCLCCDKSIYLENDVIDEKVKKIKSVTKAPSTIINNFFAGAMLGSELIKFFARDYKSMQSINSVIGIHNKNFKYEEIKLAKNYNCKYCGVNNETL

ESI-MS data for the purified protein are shown in Fig. S81.

Chemically competent E. coli BL21(DE3) cells harboring the pGro7 chaperone plasmid (Takara Bio) were transformed with pBH4-His-TEV-H. pylori MccB or pBH4-His-TEV-L. johnsonii MccB for overexpression. LB starter cultures (15 mL, 50 μg/mL carbenicillin and 25 μg/mL chloramphenicol) were grown overnight and used to inoculate 1 L LB cultures (50 μg/mL carbenicillin and 25 μg/mL chloramphenicol). Cultures were incubated at 37°C with shaking at 200 rpm until OD₆₀₀ reached 0.5–0.7. Cultures were chilled on ice for 15 minutes prior to adding 0.25 mM IPTG (isopropyl-β-D-thiogalactopyranoside) and 2 mg/L arabinose. Cultures were shaken an additional 20 hours at 18°C. Cell pellets were harvested by centrifugation at 4°C, resuspended in 40 mL lysis buffer (25 mM Tris, 500 mM NaCl, 10 mM MgCl₂, pH 8.0). Cells were lysed by three passes through an Emulsiflex microfluidizer at 15,000 psi, and the resulting lysate was centrifuged at 8,000 × g for 15 minutes at 4°C. Ni-NTA resin (1 mL) was added to the clarified lysate and His-tagged proteins were allowed to bind for 1 hour at 4°C with gentle rocking. The Ni-NTA resin was collected by centrifugation at 500 × g for 5 min, transferred to a 15 mL column, and washed with 15 mL wash buffer (20 mM Tris pH 8.0, 500 mM NaCl, 25 mM imidazole). Protein was eluted with 5 mL elution buffer (20 mM Tris pH 8.0, 500 mM NaCl, 200 mM imidazole) and dialyzed against 20 mM Tris pH 8.0, 500 mM NaCl overnight at 4°C. The dialyzed protein was then concentrated using Amicon centrifugal filter units (10,000 MWCO). Collected protein was analyzed by Coomassie-stained SDS-PAGE. Single-use aliquots were flash frozen in liquid nitrogen. Protein concentrations were determined using absorbance at 280 nm and the protein extinction coefficient as calculated using ProtParam. ESI-MS data for the purified proteins are shown in Fig. S79–S81.

Bioconjugation enzymes and protein substrates

Bioconjugation tags/epitopes are underlined.

cfGFP-EcTeCH

MGHHHHHHDYDIPTTENLYFQGSMVSKGEELFTGVVPILVELDGDVNGHKFSVSGEGEGDATYGKLTLKFISTTGKLPVPWPTLVTTLTYGVQMFARYPDHMKQHDFFKSAMPEGYVQERTIFFKDDGNYKTRAEVKFEGDTLVNRIELKGIDFKEDGNILGHKLEYNYNSHNVYITADKQKNGIKANFKIRHNIEDGGVQLADHYQQNTPIGDGPVLLPDNHYLSTQSKLSKDPNEKRDHMVLLEFVTAAGITLGMDELYKGGGGMRTGNAG

The italicized sequence was cleaved using TEV protease. ESI-MS data for the purified protein are shown in Fig. S9.

MBP-EcTeCH

MGHHHHHHDYDIPTTENLYFQGSMKIEEGKLVIWINGDKGYNGLAEVGKKFEKDTGIKVTVEHPDKLEEKFPQVAATGDGPDIIFWAHDRFGGYAQSGLLAEITPDKAFQDKLYPFTWDAVRYNGKLIAYPIAVEALSLIYNKDLLPNPPKTWEEIPALDKELKAKGKSALMFNLQEPYFTWPLIAADGGYAFKYENGKYDIKDVGVDNAGAKAGLTFLVDLIKNKHMNADTDYSIAEAAFNKGETAMTINGPWAWSNIDTSKVNYGVTVLPTFKGQPSKPFVGVLSAGINAASPNKELAKEFLENYLLTDEGLEAVNKDKPLGAVALKSYEEELAKDPRIAATMENAQKGEIMPNIPQMSAFWYAVRTAVINAASGRQTVDEALKDAQTNSSSNNNNNNNNNNLGIEGRGGGGGMRTGNAG

The italicized sequence was cleaved using TEV protease. ESI-MS data for the purified protein are shown in Fig. S14.

Protein L-EcTeCH

MHHHHHHKEETPETPETDSEEEVTIKANLIFANGSTQTAEFKGTFEKATSEAYAYADTLKKDNGEYTVDVADKGYTLNIKFAGKEKTPEEPKEEVTIKANLIYADGKTQTAEFKGTFEEATAEAYRYADALKKDNGEYTVDVADKGYTLNIKFAGKEKTPEEPKEEVTIKANLIYADGKTQTAEFKGTFEEATAEAYRYADLLAKENGKYTVDVADKGYTLNIKFAGKEKTPEEPKEEVTIKANLIYADGKTQTAEFKGTFAEATAEAYRYADLLAKENGKYTADLEDGGYTINIRFAGKKVDEKPEEKEQVTIKENIYFEDGTVQTATFKGTFAEATAEAYRYADLLSKEHGKYTADLEDGGYTINIRFAGGGGSGGGSMRTGNAG

ESI-MS data for the purified protein are shown in Fig. S20.

PTP1B(1–321)-EcTeCH

MGHHHHHHDYDIPTTENLYFQGSMEMEKEFEQIDKSGSWAAIYQDIRHEASDFPCRVAKLPKNKNRNRYRDVSPFDHSRIKLHQEDNDYINASLIKMEEAQRSYILTQGPLPNTCGHFWEMVWEQKSRGVVMLNRVMEKGSLKCAQYWPQKEEKEMIFEDTNLKLTLISEDIKSYYTVRQLELENLTTQETREILHFHYTTWPDFGVPESPASFLNFLFKVRESGSLSPEHGPVVVHCSAGIGRSGTFCLADTCLLLMDKRKDPSSVDIKKVLLEMRKFRMGLIQTADQLRFSYLAVIEGAKFIMGDSSVQDQWKELSHEDLEPPPEHIPPPPRPPKRILEPHNGGGGMRTGNAG

The italicized sequence was cleaved using TEV protease. ESI-MS data for the purified protein are shown in Fig. S16.

His-Tev-GFP-LPETGG

MGHHHHHHDYDIPTTENLYFQGSMVSKGEELFTGVVPILVELDGDVNGHKFSVSGEGEGDATYGKLTLKFISTTGKLPVPWPTLVTTLTYGVQMFARYPDHMKQHDFFKSAMPEGYVQERTIFFKDDGNYKTRAEVKFEGDTLVNRIELKGIDFKEDGNILGHKLEYNYNSHNVYITADKQKNGIKANFKIRHNIEDGGVQLADHYQQNTPIGDGPVLLPDNHYLSTQSKLSKDPNEKRDHMVLLEFVTAAGITLGMDELYKGGGGLPETGG

The italicized sequence was cleaved using TEV protease. ESI-MS data for the purified protein are shown in Fig. S55.

His-SUMO-GFP-LPETGG

MGHHHHHHDYDIPTTENLYFQGSSDSEVNQEAKPEVKPEVKPETHINLKVSDGSSEIFFKIKKTTPLRRLMEAFAKRQGKEMDSLRFLYDGIRIQADQTPEDLDMEDNDIIEAHREQIGGMVSKGEELFTGVVPILVELDGDVNGHKFSVSGEGEGDATYGKLTLKFICTTGKLPVPWPTLVTTLTYGVQCFSRYPDHMKQHDFFKSAMPEGYVQERTIFFKDDGNYKTRAEVKFEGDTLVNRIELKGIDFKEDGNILGHKLEYNYNSHNVYIMADKQKNGIKVNFKIRHNIEDGSVQLADHYQQNTPIGDGPVLLPDNHYLSTQSALSKDPNEKRDHMVLLEFVTAAGITLGMDELYKGGGGLPETGGAAASRSGC

The italicized portion was cleaved with SUMO protease Ulp1 purchased from Trialtus Bioscience according to the manufacturer’s instructions. ESI-MS data for the purified protein are shown in Fig. S59.

The appropriate plasmid was transformed into chemically competent E. coli BL21(DE3) cells for overexpression. LB starter cultures (15 mL, 50 μg/mL carbenicillin for pBH4 backbone or 50 μg/mL kanamycin for pET28a backbone) were grown overnight and used to inoculate 1 L LB cultures (50 μg/mL carbenicillin). Cultures were incubated at 37°C with vigorous shaking until OD₆₀₀ reached ~0.6. Cultures were chilled on ice for 15 minutes prior to adding 0.4 mM IPTG (isopropyl-β-D-thiogalactopyranoside). Cultures were shaken an additional 16 hours at 18°C. Cell pellets were harvested by centrifugation at 4°C, resuspended in 40 mL wash buffer (50 mM sodium phosphate, 300 mM NaCl, 20 mM imidazole, pH 8.0). Cells were lysed by three passes through an Emulsiflex microfluidizer at 15000 psi, and the resulting lysate was centrifuged at 8000 × g for 15 minutes at 4°C. Ni-NTA resin was added to the clarified lysate and allowed to bind for 1 hour at 4°C with gentle rocking. The Ni-NTA resin was collected by centrifugation at 500 × g for 5 min, transferred to a 15 mL column, and washed with 15 mL wash buffer. Protein was eluted with 5 mL elution buffer (50 mM sodium phosphate, 300 mM NaCl, 250 mM imidazole, pH 8.0) and dialyzed overnight at 4°C against wash buffer containing 1 mM DTT. To remove the His tag, TEV protease was added to the dialysis tubing at a ratio of 1:50 protease:substrate. After 16–24 hours digestion, the protein solution was passed through 2 mL Ni-NTA resin equilibrated with wash buffer. The resulting protein was then concentrated using Amicon centrifugal filter units (10,000 MWCO) and further purified by size-exclusion chromatography (Superdex Hiload 75, GE Healthcare) using storage buffer (20 mM Tris pH 8.0, 150 mM NaCl, 10% glycerol). Glycerol was omitted from the storage buffer for experiments with MccB/subtiligase. Collected fractions were analyzed by Coomassie-stained SDS-PAGE. Fractions containing the correct protein molecular weight were pooled, concentrated, and flash frozen in liquid nitrogen. Protein concentrations were determined using absorbance at 280 nm and the protein extinction coefficient as calculated using ProtParam.

GFP-TeCH-Mes thioester from GFP-TeCH-Mxe GyrA

MVSKGEELFTGVVPILVELDGDVNGHKFSVSGEGEGDATYGKLTLKFISTTGKLPVPWPTLVTTLTYGVQMFARYPDHMKQHDFFKSAMPEGYVQERTIFFKDDGNYKTRAEVKFEGDTLVNRIELKGIDFKEDGNILGHKLEYNYNSHNVYITADKQKNGIKANFKIRHNIEDGGVQLADHYQQNTPIGDGPVLLPDNHYLSTQSKLSKDPNEKRDHMVLLEFVTAAGITLGMDELYKGGGGMRTGNAGCITGDALVALPEGESVRIADIVPGARPNSDNAIDLKVLDRHGNPVLADRLFHSGEHPVYTVRTVEGLRVTGTANHPLLCLVDVAGVPTLLWKLIDEIKPGDYAVIQRSAFSVDCAGFARGKPEFAPTTYTVGVPGLVRFLEAHHRDPDAQAIADELTDGRFYYAKVASVTDAGVQPVYSLRVDTADHAFITNGFVSHATGLTGLNSGLTTNPGVSAWQVNTAYTAGQLVTYNGKTYKCLQPHTSLAGWEPSNVPALWQLQ

The italicized portion was removed by intein self-splicing. ESI-MS data for the purified protein are shown in Fig. S12.

GFP-TeCH-Mxe GyrA intein fusion protein was purified according to IMPACT Kit instruction manual (NEB #E6901S). Briefly, chemically competent E. coli BL21(DE3) cells were transformed with pTXB1-Mxe-GFP-TeCH for overexpression. LB starter culture (15 mL, 50 μg/mL carbenicillin) were grown overnight and used to inoculate 1 L LB culture (50 μg/mL carbenicillin). Culture was incubated at 37°C with vigorous shaking until OD600 reached ~0.5. Culture was chilled on ice for 15 minutes prior to the addition of 0.4 mM IPTG (isopropyl-β-D-thiogalactopyranoside). Expression was carried out at 15°C overnight to help increase the cleavage efficiency of the intein. Cell pellets were then harvested by centrifugation at 4°C, resuspended in 100 mL of ice-cold column buffer (20mM HEPES, 500mM NaCl, pH 8.5). Cells were lysed by three passes through an Emulsiflex microfluidizer at 15000 psi, and the resulting lysate was centrifuged at 8000 × g for 15 minutes at 4°C (unless otherwise noted the next steps were performed at 4°C). The clarified lysate was then loaded onto a chitin column. Before loading the crude cell extract, the chitin resin bed was washed with 10 column volumes of the column buffer (20mM HEPES, 500mM NaCl, pH 8.5). Following loading, the column was washed with 20 bed volumes of the column buffer. To induce on-column cleavage, the column was quickly flushed with 3 column volumes of cleavage buffer containing Mesna as the thiol reagent (20mM HEPES, 500mM NaCl, 50mM Mesna, pH 8.5). The flow was stopped after the quick flush and the column was incubated at 23°C for 40 hours to ensure maximum cleavage efficiency. Protein was eluted with the column buffer by resuming the flow and dialyzed against storage buffer and concentrated to appropriate concentration using Amicon centrifugal filter units (10,000 MWCO). Single-use aliquots were flash-frozen in liquid nitrogen and stored at −80°C.

Ubc9

MGHHHHHHDYDIPTTENLYFQGSMSGIALSRLAQERKAWRKDHPFGFVAVPTKNPDGTMNLMNWECAIPGKKGTPWEGGLFKLRMLFKDDYPSSPPKCKFEPPLFHPNVYPSGTVCLSILEEDKDWRPAITIKQILLGIQELLNEPNIQDPAQAEAYTIYCQNRVEYEKRVRAQAKKFAPS

ESI-MS data for the purified protein are shown in Fig. S82.

GFP-LACE(D173)

MGHHHHHHDYDIPTTENLYFQGMRKGEELFTGVVPILVELDGDVNGHKFSVRGEGEGDATNGKLTLKFICTTGKLPVPWPTLVTTLTYGVQCFARYPDHMKQHDFFKSAMPEGYVQERTISFKDDGTYKTRAEVKFEGDTLVNRIELKGIDFKEDGNILGHKLEYNFNSHNVYITADKQKNGIKANFKIRHNVEDGSGPRKVIKMESEEGSGSVQLADHYQQNTPIGDGPVLLPDNHYLSTQSVLSKDPNEKRDHMVLLEFVTAAGITHGMDELYK

The italicized sequence was cleaved using TEV protease. ESI-MS data for the purified protein are shown in Fig. S74.

The appropriate plasmid was transformed into chemically competent E. coli BL21(DE3) cells for overexpression. LB starter cultures (15 mL, 50 μg/mL carbenicillin for pBH4 backbone or 50 μg/mL kanamycin for pET28a backbone) were grown overnight and used to inoculate 1 L LB cultures (50 μg/mL carbenicillin). Cultures were incubated at 37°C with vigorous shaking until OD₆₀₀ reached ~0.6. Cultures were chilled on ice for 15 minutes prior to adding 0.4 mM IPTG (isopropyl-β-D-thiogalactopyranoside). Cultures were shaken an additional 16 hours at 18°C (GFP-LACE) or 4 h at 30°C (Ubc9). Cell pellets were harvested by centrifugation at 4°C, resuspended in 40 mL wash buffer (50 mM HEPES, 350 mM NaCl, 20 mM imidazole, pH 8.0). Cells were lysed by three passes through an Emulsiflex microfluidizer at 15000 psi, and the resulting lysate was centrifuged at 8000 × g for 15 minutes at 4°C. Ni-NTA resin was added to the clarified lysate and allowed to bind for 1 hour at 4°C with gentle rocking. The Ni-NTA resin was collected by centrifugation at 500 × g for 5 min, transferred to a 15 mL column, and washed with 15 mL wash buffer. Protein was eluted with 5 mL elution buffer (50 mM sodium phosphate, 300 mM NaCl, 250 mM imidazole, pH 8.0) and dialyzed overnight at 4°C against wash buffer. To remove the His tag from GFP-LACE, TEV protease was added to the dialysis tubing at a ratio of 1:50 protease:substrate. After 16–24 hours digestion, the protein solution was passed through 2 mL Ni-NTA resin equilibrated with wash buffer. The resulting protein was then concentrated using Amicon centrifugal filter units (10,000 MWCO). Glycerol was omitted from the storage buffer for experiments with MccB/subtiligase. Purified protein was analyzed by Coomassie-stained SDS-PAGE. Protein concentrations were determined using absorbance at 280 nm and the protein extinction coefficient as calculated using ProtParam. Single-use aliquots were flash frozen in liquid nitrogen and stored at −80°C.

Anti-GFP rAb Heavy Chain-TeCH

Light chain:

MKSLLPTAAAGLLLLAAQPAMASDIQMTQSPSSLSASVGDRVTITCRASQSVSSAVAWYQQKPGKAPKLLIYSASSLYSGVPSRFSGSRSGTDFTLTISSLQPEDFATYYCQQSWGLITFGQGTKVEIKRTVAAPSVFIFPPSDSQLKSGTASVVCLLNNFYPREAKVQWKVDNALQSGNSQESVTEQDSKDSTYSLSSTLTLSKADYEKHKVYACEVTHQGLSSPVTKSFNRGEC

Heavy chain:

MKKNIAFLLASMFVFSIATNAYAEISEVQLVESGGGLVQPGGSLRLSCAASGFNISYYSIHWVRQAPGKGLEWVASIYPYYSSTSYADSVKGRFTISADTSKNTAYLQMNSLRAEDTAVYYCARAGWVASSGMDYWGQGTLVTVSSASTKGPSVFPLAPSSKSTSGGTAALGCLVKDYFPEPVTVSWNSGALTSGVHTFPAVLQSSGLYSLSSVVTVPSSSLGTQTYICNVNHKPSNTKVDKKVEPKSCGGGGGMRTGNAG

ESI-MS data for the purified protein are shown in Fig. S18.

pPSL937-anti-GFP⁶⁷ rAb-HC-TeCH was transformed into chemically competent C43(DE3) Pro+ pTUM protease-deficient E. coli cells for overexpression. LB starter cultures (5 mL, 50 μg/mL carbenicillin, 25 μg/mL chloramphenicol) were grown overnight and used to inoculate 1 L TB expression cultures (50 μg/mL carbenicillin, 25 μg/mL chloramphenicol, 0.05% w/v glucose, 0.5% w/v lactose, 1% w/v galactose, 2 mM MgSO₄). Cultures were incubated at 37°C for 6 hours and grown for another 18 hours at 30°C. Cell pellets were harvested by centrifugation at 4°C and resuspended in 40 mL 1:1 PBS:Bacterial Protein Extraction Reagent (BPER, ThermoFisher Scientific). Cell pellets were lysed by incubating in 60°C water bath for 30 minutes. The resulting lysate was centrifuged at 8,000 × g for 15 minutes at 4°C. The supernatant was filtered with 0.2 μm syringe filters and loaded onto a HiTrap Protein A HP column (Cytiva). The column was washed with PBS and protein was eluted with 100 mM acetic acid and immediately neutralized. Fractions were analyzed by Coomassie-stained SDS-PAGE. Fractions containing the correct protein molecular weight were pooled, concentrated, and flash frozen in liquid nitrogen. Protein concentrations were determined using absorbance at 280 nm and the protein extinction coefficient as calculated using ProtParam.

Subtiligase

MRGKKVWISLLFALALIFTMAFGSTSSAQAAGKSNGEKKYIVGFKQTMSTMSAAKKKDVISEKGGKVQKQFKYVDAASATLNEKAVKELKKDPSVAYVEEDHVAHAYAQSVPYGVSQIKAPALHSQGYTGSNVKVAVIDSGIDSSHPDLKVAGGASFVPSETNPFQDNNSHGTHVAGTVAALDNSIGVLGVAPSASLYAVKVLGADGSGQYSWIISGIEWAIANNMDVINLALGGPSGSAALKAAVDKAVASGVVVVAAAGNEGTSGSSSTVGYPGKYPSVIAVGAVDSSNQRASFSSVGPELDVMAPGVSIQSTLPGNRYGAYSGTCMASAHVAGAAALILSKHPNWTNTQVRSSLENTTTKLGDSFYYGKGLINVQAAAQLEHHHHHH

The italicized portion of the protein is removed by autoproteolysis. ESI-MS data for the purified protein are shown in Fig. S83.

B. subtilis BG2864 cell were transformed with pBS42-pre-pro-stabiligase-His₆ as described previously⁵¹.2xYT starter cultures (15 mL, 12.5 μg/mL chloramphenicol) were grown overnight and used to inoculate 200 mL 2xYT cultures (12.5 μg/mL chloramphenicol, 5 mM CaCl₂) to OD₆₀₀ 0.03–0.05. Cultures were incubated at 37°C with vigorous shaking for 24 hours. Cells were pelleted by centrifugation at 4°C, and the resulting supernatant was added to 3 volumes of ice-cold ethanol. The precipitate was harvested by centrifugation at 4°C and resuspended in wash buffer (50 mM sodium phosphate, 300 mM NaCl, 20 mM imidazole, pH 8.0). After brief centrifugation, Ni-NTA resin (400 μL) was added to the resuspended pellet and His-tagged protein was allowed to bind for 1 hour at 4°C with gentle rocking. The Ni-NTA resin was collected by centrifugation at 500 × g for 5 min, transferred to a 1 mL spin column, and washed with 4 mL wash buffer. Protein was eluted with 0.8 mL elution buffer (50 mM sodium phosphate, 300 mM NaCl, 250 mM imidazole, pH 8.0). The resulting protein was then concentrated using Amicon centrifugal filter units (10,000 MWCO) and further purified by size-exclusion chromatography (Superdex Hiload 75, GE Healthcare) using storage buffer (100 mM bicine, 5 mM DTT, pH 8.5). Collected fractions were analyzed by Coomassie-stained SDS-PAGE. Fractions containing the correct protein molecular weight were pooled, concentrated, and flash frozen in liquid nitrogen. Protein concentrations were determined using absorbance at 280 nm and the protein extinction coefficient as calculated using ProtParam.

eSrtA

MQAKPQIPKDKSKVAGYIEIPDADIKEPVYPGPATREQLNRGVSFAEENESLDDQNISIAGHTFIDRPNYQFTNLKAAKKGSMVYFKVGNETRKYKMTSIRNVKPTAVEVLDEQKGKDKQLTLITCDDYNEETGVWETRKIFVATEVKLEHHHHHH

ESI-MS data for the purified protein are shown in Fig. S84.

Chemically competent E. coli BL21(DE3) cells were transformed with pET29a-eSrta and plated on LB-agar containing kanamycin (50 μg/mL). A starter culture was prepared by inoculating LB media (10 mL) containing kanamycin (50 μg/mL) with a single colony. The culture was grown overnight at 37°C with shaking at 200 rpm. The starter culture was used to inoculate 1 L LB media in a baffled flask for overexpression. The culture was incubated at 37°C with shaking at 200 rpm until OD₆₀₀ reached ~0.6. IPTG was then added to a final concentration of 0.4 mM and the culture was incubate for an additional 3 h at 30°C for protein expression. Cultures were centrifuged at 4,000 × g at 4°C for 20 min to pellet the cells. Cells were resuspended in 50 mL lysis buffer (50 mM Tris pH 8.0, 300 mM NaCl supplemented with 100 μM PMSF and Roche Complete EDTA-free Protease Inhibitor) and were lysed by sonication using a QSonica 700 probe sonicator (50% amplitude, 5 s on / 5 s off, 15 cycles). Cell debris was pelleted by centrifugation at 10,000 × g for 20 minutes at 4°C. The resultant supernatant was added to a conical tube containing 1 mL of Qiagen Ni-NTA agarose and the mixture was incubated for 1 hour at 4°C with gentle rocking. The Ni-NTA resin was collected by centrifugation at 500 × g for 5 min, transferred to a 15 mL column, and washed with 10 mL wash buffer (50 mM Tris pH 8.0, 300 mM NaCl, 20 mM imidazole). Protein was eluted with 5 mL elution buffer (50 mM Tris pH 8.0, 300 mM NaCl, 250 mM imidazole) and dialyzed against storage buffer (25 mM Tris pH 7.5, 150 mM NaCl) overnight at 4°C. The dialyzed protein was then concentrated using Amicon centrifugal filter units (10,000 MWCO). Protein concentrations were determined using absorbance at 280 nm and the protein extinction coefficient as calculated using ProtParam. Single-use aliquots were flash frozen in liquid nitrogen and stored at −80°C until further use.

Sequence similarity network construction.

A sequence similarity network (SSN) for representative members of the E1-like/ThiF superfamily was constructed using the Enzyme Function Initiative Enzyme Similarity Tool^68,69 at https://efi.igb.illinois.edu/efi-est/. Input sequences consisted of those retrieved from the ThiF family in Pfam⁷⁰ and were comprised of reviewed Uniprot⁷¹ entries and MccB sequences with annotated MccA substrates from a bioinformatic analysis of MccB homologs involved in biosynthesis of microcin C³² . The SSN was visualized in Cytoscape 3.9.1 using a minimum alignment score of 100.

Purine nucleoside phosphorylase (PNP)/inorganic pyrophosphate (IP)-coupled kinetic assays.

The kinetics of MccB-catalyzed adenylation were characterized using an enzyme-coupled assay for detecting formation of pyrophosphate EnzCheck Pyrophosphate Assay Kit (Thermo Fisher Scientific) (Fig. S85). To screen MccB for adenylation activity with C-terminal variants of MccA, 5 μM of the appropriate MccB variant was incubated with 0.25 mM of each MccA variant in a reaction with 0.25 mM ATP, 5 mM MgCl₂, 0.5 U/mL purine nucleoside phosphorylase (PNP), and 0.5 U/mL inorganic pyrophosphatase (IP). To screen E. coli, H. pylori, and L. johnsonii MccB homologs for adenylation activity with MccA homologs, 5 μM of the appropriate MccB enzyme was incubated with 0.25 mM of each MccA variant in a reaction with 0.25 mM ATP, 5 mM MgCl2, 0.5 U/mL purine nucleoside phosphorylase (PNP), and 0.5 U/mL inorganic pyrophosphatase (IP). Reactions containing HpMccA-N7G and LjMccA-N7G peptides also contained 5 mM Mesna and 5 mM TCEP. To collect steady-state kinetics data for individual peptide substrates, the peptide concentration was varied from 0–600 μM in a reaction with 5 μM MccB, 0.25 mM ATP, 5 mM MgCl₂, 0.5 U/mL purine nucleoside phosphorylase (PNP), and 0.5 U/mL inorganic pyrophosphatase (IP). Initial fitting of kinetic data for the MccA-N7G substrate revealed that the apparent K_M (8.9 ± 0.9 μM, Fig. S2) for MccA-N7G was close to the concentration of MccB (5 μM) used in the reaction, invalidating the assumption that the concentration of free substrate unbound to enzyme is approximately equal to the total substrate concentration (the free ligand approximation) that is used to derive the Michaelis-Menten equation. We therefore repeated measurements for the MccA-N7G peptide using an MccB concentration of 0.5 μM and a substrate concentration range from 0–100 μM. Reactions were initiated with the addition of MccB and absorbance at 360 nm was monitored using a Tecan Infinite M200 plate reader. Initial rates of pyrophosphate production were calculated based on the initial rate of absorbance change and the resulting purine analog extinction coefficient (11,000 M⁻¹cm⁻¹). Reactions were carried out in triplicate and data analysis was performed using GraphPad Prism 10.

LC-TOF MS analysis of MccB reactions with synthetic peptide substrates.

For analysis of formation of N-AMPylated MccA, MccA substrate (250 μM) was incubated with 5 μM MccB and 5 mM ATP in reaction buffer (75 mM Tris pH 8.0, 5 mM MgCl₂). Reactions were incubated for 16 hours at room temperature and quenched with addition of an equivalent volume of 0.6% TFA. These reaction conditions were used for MccBs from E. coli, H. pylori, L. johnsonii, and H. somni. Quenched reactions were centrifuged at 8,000 × g for 10 minutes and analyzed using LC-TOF MS as described in Mass spectrometry.

For analysis of E. coli MccB-catalyzed thioesterification, MccA-N7G substrate (250 μM) was incubated with 5 μM MccB, 5 mM ATP, 5 mM Mesna, and 5 mM TCEP in reaction buffer (75 mM Tris pH 8.0, 5 mM MgCl₂). Reactions were incubated for 16 hours at room temperature and quenched with addition of an equivalent volume of 0.6% TFA. To screen the ability of E. coli MccB, H. pylori MccB, L. johnsonii MccB to catalyze thioesterification of MccA-N7Gs derived from each species, reactions contained 5 μM MccB, 5 mM ATP, 50 mM Mesna, and 12.5 mM TCEP in reaction buffer (75 mM Tris pH 8.0, 5 mM MgCl₂). To screen the ability of H. somni MccB to catalyze thioesterification of HsLACE peptides, reactions contained MccB (5–25 μM), ATP (5 mM), MgCl₂ (5 mM), peptide substrate (0.25 mM), 25 mM Mesna or AcCysNHMe, and reaction buffer (100 mM HEPES, pH 8.0). Quenched reactions were centrifuged at 8,000 × g for 10 minutes and analyzed using LC-TOF MS as described in Mass spectrometry.

Screening nucleophiles for MccB-mediated ligation of synthetic peptide substrates.

Nucleophile were dissolved in water and adjusted to neutral pH using paper pH strips (EMD Millipore). Solutions were stored at −80°C until further use, unless otherwise noted. MccA-N7G (250 μM) was incubated with 5 μM MccB, 5 mM ATP, and the corresponding concentration of nucleophile in reaction buffer (25 mM Tris pH 8.0, 10 mM MgCl₂). Reactions were incubated for 16 hours at room temperature and quenched with addition of an equivalent volume of 0.6% TFA. Quenched reactions were centrifuged at 8,000 × g for 10 minutes and analyzed using LC-TOF MS as described in Mass spectrometry. Extracted ion chromatogram (EIC) peak areas were calculated using Agilent TOF Quantitative Analysis v11.0. Percent conversion of substrate to nucleophile-modified product was calculated using the integrated peak areas from the substrate peptide EIC and product ion EIC using the formula shown below.

$$\mathit{percent}\mathit{conversion}=\frac{\mathit{product}\mathit{peak}\mathit{area}}{\mathit{product}\mathit{peak}\mathit{area}+\mathit{substrate}\mathit{peak}\mathit{area}}\times 100$$

Raw data are available in the Dryad repository at DOI: 10.5061/dryad.c59zw3rkb⁶⁶.

MccB-mediated C-terminal ligation of MccA-N7G with Cys.

MccA-N7G (250 μM) was incubated for 16 hours at room temperature in reactions containing 5 μM MccB, 5 mM ATP, and the appropriate concentration of Cys/TCEP in reaction buffer (25 mM Tris pH 8.0, 10 mM MgCl₂). For maleimide functionalization, reactions were performed with 1 mM Cys/TCEP for 16 hours followed by a 1-hour incubation with 1.5 mM maleimide. All reactions were quenched, centrifuged at 8,000 × g for 10 minutes, and analyzed using LC-TOF MS as described in Mass spectrometry.

MccA-N7G thioester exchange with N-terminal Cys peptide.

C-terminal thioester was prepared in reactions containing 250 μM MccA-N7G, 5 μM MccB, 5 mM ATP, and 5 mM Mesna/TCEP in reaction buffer (25 mM Tris pH 8.0, 10 mM MgCl₂). After incubating for 16 hours at room temperature, 20 mM Cys-Trp peptide was added to each reaction and incubated for an additional 4 hours. Reactions were quenched with the addition of an equivalent volume of 0.6% TFA. Quenched reactions were centrifuged at 8000 × g for 10 minutes and analyzed using LC-TOF MS as described in Mass spectrometry.

MccB-mediated ligation of GFP-MccA-N7G protein substrates.

GFP-MccA-N7G variant (50 μM) was incubated with 5 μM MccB, 5 mM ATP, and the appropriate concentration of cysteine or Mesna/TCEP in reaction buffer (25 mM Tris pH 8.0, 10 mM MgCl₂). Reactions were incubated at room temperature for the indicated times. For labeling time course experiments, reactions were quenched with the addition of an equivalent volume of 0.6% TFA. For maleimide functionalization of Cys-ligated proteins, reactions were desalted using 75 μL 7K MWCO Zeba Micro Spin Desalting Columns (Thermo Fisher Scientific) and incubated with maleimides overnight at 4°C. Reactions were centrifuged at 8,000 × g for 10 minutes and analyzed using LC-TOF MS as described in Mass spectrometry.

GFP-MccA-N7G thioester exchange time course with Cys-Trp dipeptide.

C-terminal thioester was prepared in reactions containing 50 μM cysteine-free GFP-MccA-N7G, 5 μM MccB, 5 mM ATP, and 5 mM Mesna/TCEP in reaction buffer (25 mM Tris pH 8.0, 10 mM MgCl₂). After incubating for 1 h at room temperature reactions were desalted using 75 μL 7K MWCO Zeba Micro Spin Desalting Columns (Thermo Fisher Scientific). Cys-Trp was added to a final concentration of 5 mM. Reactions were quenched at various time points by addition of an equivalent volume of 0.6% TFA. Quenched reactions were centrifuged at 8,000 × g for 10 minutes and analyzed using LC-TOF MS as described in Mass spectrometry.

GFP-MccA-N7G expressed protein ligation with N-terminal Cys peptide followed by copper-free click chemistry.

Reactions contained 50 μM GFP-MccA-N7G, 5 μM MccB, 5 mM ATP, 5 mM Mesna/DTT, and 5 mM CGAGS-azidoAla peptide in reaction buffer (25 mM Tris pH 8.0, 10 mM MgCl₂). After incubating for 4 h at room temperature, reactions were desalted using 75 μL 7K MWCO Zeba Micro Spin Desalting Columns (Thermo Fisher Scientific). Desalting was repeated twice more. Dibenzocyclooctyne (DBCO)-PEG4-biotin was added to a final concentration of 3 mM and incubated for 1 h at room temperature. Reactions were desalted once more and DTT was added to a final concentration of 5 mM for a 30-minute incubation at room temperature. Reactions were centrifuged at 8,000 × g for 10 minutes and analyzed using LC-TOF MS as described in Mass spectrometry.

LC-TOF MS analysis of MccB-catalyzed N-AMPylation of MccA positional scanning peptide library.

MccA variants were synthesized individually in-house (according to Solid-phase peptide synthesis) or were purchased from Peptide2.0. For analysis of formation of N-AMPylated MccA variants, MccA substrate (250 μM) was incubated with 5 μM MccB and 5 mM ATP in reaction buffer (75 mM Tris pH 8.0, 5 mM MgCl₂). Reactions were incubated for 16 hours at room temperature and quenched with addition of an equivalent volume of 0.6% TFA. Quenched reactions were centrifuged at 8,000 × g for 10 minutes and analyzed using LC-TOF MS as described in Mass spectrometry. Percent conversion was calculated using peak areas from substrate and product EICs. Raw data are available in the Dryad repository at DOI: 10.5061/dryad.c59zw3rkb⁶⁶.

MccB homolog reactions with GFP-TeCH-tag fusion proteins.

E. coli, H. pylori, and L. johnsonii were screened for protein thioesterification using GFP fused with the MccA-N7G homolog sequences. Reactions contained 5 μM MccB, 5 mM ATP, 50 μM GFP-TeCH fusion, 50 mM Mesna, and 12.5 mM TCEP in reaction buffer (75 mM Tris-HCl pH 8.0, 5 mM MgCl₂). For multiplexed reactions containing all three GFP-TeCH fusions, each GFP-TeCH fusion was added to a final concentration of 16 μM. Reactions were quenched after 16 hours at room temperature by adding two volumes of 0.6% TFA. Samples were analyzed using LC-TOF MS as described in Mass spectrometry.

MccB/subtiligase-catalyzed ligation of protein substrates.

MccB/subtiligase-catalyzed C-terminal amide bond formation was performed in one-pot reactions containing 50 μM TeCH-tagged protein, 5 μM MccB, 5 mM ATP, 5 μM subtiligase, 5 mM peptide substrate, and 5 mM Mesna/DTT in reaction buffer (25 mM Tris pH 8.0, 10 mM MgCl₂). After incubating for appropriate time at room temperature, reactions were diluted in 75 mM HEPES pH 8.0 and analyzed using LC-TOF MS as described in Mass spectrometry.

MccB/subtiligase-mediated ligation of anti-GFP rAb followed by copper free click chemistry.

C-terminal bioconjugation was performed in one-pot reactions containing 50 μM protein, 5 μM MccB, 5 mM ATP, 5 μM subtiligase, 5 mM peptide substrate, and 5 mM Mesna in reaction buffer (75 mM Tris pH 8.0, 5 mM MgCl₂). After incubating for the appropriate time at room temperature, reactions were twice buffer exchanged into 75 mM HEPES, pH 8.0 using 500 μL 7K MWCO Zeba Micro Spin Desalting Columns (Thermo Fisher Scientific), and analyzed using LC-TOF MS. Absorbance at 280 nm was used to estimate the resulting protein concentration, and 2–5 molar equivalents of the DBCO reagent were added to each sample. After 4 hours at room temperature, the samples were twice buffer exchanged as before, and analyzed using LC-TOF MS as described in Mass spectrometry.

Cell culture and immunofluorescence.

Flp-In T-Rex 293T cells (Thermo Fisher Scientific) were grown in DMEM supplemented with 10% fetal bovine serum, 100 U/mL penicillin, 100 μg/mL streptomycin, and other antibiotics as appropriate. Cells were tested every six months for mycoplasma contamination using the LookOut Mycoplasma PCR Detection Kit (Sigma-Aldrich) according to the manufacturer’s instructions. A stable cell line expressing Igκ–eGFP-PDGF receptor β-chain transmembrane domain (PDGFRTM) under a doxycycline-inducible promoter was introduced by transfecting cells with pcDNA5/FRT/TO-Igκ-eGFP-PDGFRTM according to the manufacturer’s instructions. This construct included an N-terminal Igκ signal peptide and a C-terminal PDGF receptor β-chain transmembrane domain to target eGFP to the cell surface. For immunofluorescence experiments, cells were seeded at 10,000 cells per well in a 96-well plate and grown to 50% confluency. Doxycycline (1 ug/mL) was then added to induce cell surface GFP expression. After 20 hours, cells were washed three times with ice cold PBS, fixed with PBS + 4% paraformaldehyde for 10 minutes, and washed three times with PBS + 3% BSA. Cells were stained with 10 μg/mL αGFP-rAb in PBS + 3% BSA for 1 hour at room temperature and washed three times with PBS + 3% BSA. Imaging was performed on an Echo Revolve epifluorescence microscope in the inverted configuration.

Igκ-GFP-PDGFR-TM sequence.

The N-terminal Igκ leader sequence is underlined, a V5 tag is shown in italic, and the C-terminal PDGF receptor transmembrane domain is underlined.

METDTLLLWVLLLWVPGSTGGKPIPNPLLGLDSTGSGGGASMVSKGEELFTGVVPILVELDGDVNGHKFSVSGEGEGDATYGKLTLKFICTTGKLPVPWPTLVTTLTYGVQCFSRYPDHMKQHDFFKSAMPEGYVQERTIFFKDDGNYKTRAEVKFEGDTLVNRIELKGIDFKEDGNILGHKLEYNYNSHNVYIMADKQKNGIKVNFKIRHNIEDGSVQLADHYQQNTPIGDGPVLLPDNHYLSTQSALSKDPNEKRDHMVLLEFVTAAGITLGMDELYKGSGGSGGGGSAVGQDTQEVIVVPHSLPFKVVVISAILALVVLTIISLIILIMLWQKKPR

C-terminal labeling using eSrtA.

GFP-LPETGG (50 μM) was incubated with eSrtA (2.5 μM) and a triglycine peptide (GGG, variable concentration) in reaction buffer (50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 10 mM CaCl₂). Reactions were incubated at room temperature for 30 minutes and then analyzed by LC-TOF MS as described in Mass spectrometry. Percent conversion of GFP-LPETGG (27,814 Da) to cyclized GFP (27,681 Da) or ligated GFP (27,871 Da) was calculated using peak areas from the deconvoluted mass spectra.

Dual N- and C-terminal labeling with eSrtA and MccB/subtiligase.

For one-pot reactions, Gly-Ser-MBP-TeCH (25 μM) was incubated with MccB (5 μM), subtiligase (5 μM), and AFAGAGS-azidoAla (5 mM, MccB/subtiligase substrate), 5 mM ATP, eSrtA (2.5 μM), and 5-FAM-LPETGG (2 mM) in reaction buffer (100 mM HEPES, pH 8.0, 150 mM NaCl, 10 mM CaCl₂, 5 mM MgCl₂) for 12 h. For telescoping reactions, Gly-Ser-MBP-TeCH (25 μM) was incubated with MccB (5 μM), subtiligase (5 μM), and AFAGAGS-azidoAla (5 mM, MccB/subtiligase substrate) in reaction buffer for 4 h at room temperature, followed by addition of eSrtA (2.5 μM) and 5-FAM-LPETGG (2 mM) and incubation for an additional 4 h at room temperature.

Chemical synthesis of LRLRGG-Mes thioester peptide.

LRLRGG-hydrazide was purchased from GenScript. The hydrazide peptide was converted to the Mes thioester as previously described^49,72 . LRLRGG-hydrazine (4 mg, 6 μmol, 1 equivalent) was dissolved in 200 mM sodium phosphate, 6 M guanidinium hydrochloride (GdnHCl, 0.5 mL). The solution was cooled with stirring to −15°C in a bath of aqueous saturated sodium chloride and ice. A freshly prepared aqueous solution of sodium nitrite (0.5 M, 0.12 mL, 60 μmol, 10 equivalents) was added and the reaction was allowed to proceed for 20 min. A solution of sodium mercaptoethanesulfonate (Mesna, 29.6 mg, 180 μmol, 30 equivalents) in 200 mM sodium phosphate, pH 7.0, 6 M GdnHCl (0.72 mL) was added to the reaction. The mixture was adjusted to pH 7 using 1 N NaOH and reaction was allowed to proceed for 2 h at ambient temperature. The reaction mixture was purified using a semi-preparative ZORBAX Eclipse XDB-C18 column (9.4 × 250 mm, 5 μm) coupled to an Agilent 1260 Infinity II HPLC. A 90-min gradient from 100% mobile phase A (0.1% trifluoroacetic acid in water) to 100% B (acetonitrile) at 2 mL/min was used to separate the desired product. Fractions were analyzed by LC-TOF MS and those containing the pure product were pooled and lyophilized. The resultant product was dissolved in water and quantified using the Pierce Quantitative Fluorometric Peptide Assay Kit (Thermo Fisher Scientific). The yield was 80 μL of a 34 mM solution (2.2 mg, 2.7 μmol, 45%).

H. somni MccB-catalyzed peptide thioester synthesis.

For screening the activity of HsMccB on HsTeCH, HsLACE1–4, and LACE substrate peptide, 20 μL reactions were prepared that contained 100 mM HEPES, pH 8.0, HsMccB (5–25 μM), HsTeCH, HsLACE1–4, or LACE peptide (250 μM), ATP (5 mM), MgCl₂ (5 mM), and thiol (Mesna or AcCysNHMe, 25 mM). Reactions were initiated by addition of HsMccB and incubated for 16–20 h at room temperature. Raw data are available in the repository at DOI: 10.5061/dryad.c59zw3rkb⁶⁶.

For larger scale synthesis of MLGLRGG-Mes (HsLACE3-Mes), identical reaction conditions and reagent concentrations were used in a volume of 200 μL. Reactions were quenched by addition of trifluoroacetic acid (TFA) to 1% final concentration and were desalted on SOLA HRP C18 solid-phase extraction columns (10 mg format, Thermo Fisher Scientific). Columns were conditioned with 100% acetonitrile (500 μL) and equilibrated with 0.1% TFA (2 × 1 mL). The acidified sample was then loaded onto the column. The column was washed with 0.1% TFA (2 × 1 mL) and eluted with 80% acetonitrile/20% water (2 × 150 μL). The eluted peptide thioester was dried in a vacuum concentrator (SpeedVac SPD130DLX, Thermo Fisher Scientific) and dissolved in water. The recovered peptide was quantified using the Pierce Quantitative Fluorometric Peptide Assay Kit (Thermo Fisher Scientific). Typical recovery from desalting was 50%.

Lysine acylation using conjugating enzymes (LACE) of GFP-LACE-tag using MccB-generated thioesters.

GFP-LACE-tag⁴⁹ (15 μM), Ubc9 (60 μM), MLGLRGG-Mes or other C-terminal thioesters as indicated in the text (150 μM-1500 μM) were incubated in reaction buffer (100 mM HEPES, pH 8.0, 50 mM KCl) at 30°C for 16 h. For pH optimization, reaction buffer was either 100 mM HEPES, pH 7.6, 50 mM KCl or 100 mM HEPES, pH 8.0, 50 mM KCl. For analyzing the effect of thiol concentration on reaction efficiency, Mesna (0–25 mM) was included in the reaction. Optimal reaction conditions for transfer of MLGLRGG-Mes to GFP LACE used 100 mM HEPES, pH 8.0, 50 mM KCl and omitted thiol.

Extended Data

Extended Data Figure 1. MccB-generated thioesters undergo transthioesterification, S-to-N acyl transfer, and native chemical ligation.

(A) MccA-N7G thioester can undergo transthioesterification with MPAA, a thiol nucleophile that cannot directly capture Mcc-N7G-O-AMP. (B) MccA-N7G-O-AMP can undergo thioesterification and S-to-N acyl shift with Cys to form a peptide bond. (C) In the presence of ATP and Mesna, MccB catalyzes native chemical ligation between an unactivated peptide and an N-terminal Cys peptide.

Extended Data Figure 2. MccB homologs from L. johnsonii and H. pylori can be used in combination with subtiligase for enzyme-catalyzed expressed protein ligation.

(A) LjMccB- and subtiligase-catalyzed ATP-dependent peptide ligation of Ala-Phe (left) or AFAGAGS-azAla (right) to GFP-LjTeCH. (B) HpMccB- and subtiligase-catalyzed ATP-dependent peptide ligation of Ala-Phe (left) or AFAGAGS-azAla (right) to GFP-HpTeCH.

Extended Data Figure 3. Comparison of MccB/subtiligase-catalyzed and eSrtA-catalyzed C-terminal protein modification.

(A) MccB/subtiligase-catalyzed C-terminal peptide ligation to GS-GFP-TeCH. In the absence of peptide nucleophile, MccB and subtiligase catalyze GS-GFP cyclization, but this reaction is efficiently suppressed in the presence of 5 mM Ala-Phe. (B) eSrtA-catalyzed C-terminal modification of GS-GFP-LPETGG. In the absence of nucleophile, eSrtA catalyzes GFP cyclization that cannot be completely suppressed even in the presence of 10 mM GGG peptide. (C) eSrtA cyclization is suppressed by removing the N-terminal GS sequence at the N terminus of GS-GFP-LPETGG.

Extended Data Figure 4. Dual N- and C-terminal labeling of GS-MBP-TeCH using eSrtA and MccB/subtiligase.

(A) Scheme for dual N- and C-terminal label of GS-MBP-TeCH with MccB/subtiligase and eSrtA. The magenta circle represent azidoAla and the cyan circle represents 5-FAM. (B) Telescoping one-pot dual labeling of GS-MBP-TeCH with eSrtA and MccB/subtiligase. (C) Concurrent one-pot dual labeling of GS-MBP-TeCH with eSrtA and MccB/subtiligase.

Extended Data Figure 5. Combining HsMccB-catalyzed peptide thioester synthesis with Ubc9-catalyzed lysine acylation.

(A) Scheme for lysine acylation using an HsMccB-generated thioester and GFP with an internal minimal LACE tag sequence (IKQE). (B) Scheme for lysine acylation using an HsMccB-generated thioester and GFP with a full length LACE tag sequence (PRKVIKMESEE). (C) Optimization of peptide thioester concentration in LACE reactions. (D) Optimization of thiol concentration at pH 7.6. Excess thiol suppresses the LACE reaction. (E) Optimization of thiol concentration at pH 8.0. Excess thiol suppresses the LACE reaction, which proceeds to higher yield at pH 8.0 compared to 7.6.

Data availability

All data supporting the conclusions of this manuscript can be found in the main text, Supplementary Information, or in the Dryad repository at the URL: https://doi.org/10.5061/dryad.c59zw3rkb⁶⁶.