Engineering Bifunctional Enzymes Capable of Adenylating and

Feb 2, 2018 - Department of Pharmaceutical Sciences, College of Pharmacy, University of Kentucky, Lexington, Kentucky 40536-0596, United States. ACS S...
3 downloads 8 Views 1019KB Size
Download PDF
Letter Cite This: ACS Synth. Biol. XXXX, XXX, XXX−XXX

pubs.acs.org/synthbio

Engineering Bifunctional Enzymes Capable of Adenylating and Selectively Methylating the Side Chain or Core of Amino Acids Taylor A. Lundy, Shogo Mori, and Sylvie Garneau-Tsodikova* Department of Pharmaceutical Sciences, College of Pharmacy, University of Kentucky, Lexington, Kentucky 40536-0596, United States S Supporting Information *

ABSTRACT: Nonribosomal peptides (NRPs) are known sources of therapeutics. Some nonribosomal peptide synthetase assembly lines contain unique functional interrupted adenylation (A) domains, where nature has combined two different functional domains into one bifunctional enzyme. Most often these interrupted A domains contain a part of a methylation (M) domain embedded in their sequence. Herein, we aimed to emulate nature and create fully functional interrupted A domains by inserting two different noncognate M domains, KtzH(MH) and TioS(M3S), into a naturally occurring uninterrupted A domain, Ecm6(A1T1). We evaluated the engineered enzymes, Ecm6(A1aMHA1bT1) and Ecm6(A1aM3SA1bT1), by a series of radiometric assays and found that not only do they maintain A domain activity, but also they gain the site-specific methylation patterns observed in the parent M domain donors. These findings provide an exciting proof-of-concept for generating interrupted A domains as future tools to modify NRPs and increase the diversity and activity of potential therapeutics. KEYWORDS: echinomycin, kutznerides, natural products, nonribosomal peptide biosynthesis, synthetic biology, thiocoraline

N

responsible for activating specific amino acid-like building blocks through the process of adenylation. The amino acid building blocks are shuttled from one end of the assembly line to the other via T domains, which must first be converted from their inactive (apo) to their active (holo) state by the transfer of a 4′-phosphopantetheine (Ppant) prosthetic group from coenzyme A (CoA), catalyzed by a 4′-phosphopantetheinyltransferase (PPTase).6 The C domains catalyze peptide bond formation between two amino acid building blocks tethered to the T domains that surround them.4 The selectivity for particular substrates is dictated by A domains, which are in fact responsible for the structural diversity seen in NRPs.7 The amino acid sequences of A domains are characterized by ten conserved sequence motifs (a1−a10). These domains can be divided into the large Nterminus (a1−a7) and the small C-terminus (a8−a10) subunits.8 A domains have been the focus of many genetic engineering studies9 with the aim of changing or expanding their substrate specificity. A few approaches have produced derivatized NPs. These approaches include changing the residues of the substrate binding pocket10,11 or replacing an entire A domain in a pathway with a different one.12 These methods, however, can result in substantial loss of adenylating activity or disruption of domain communication, stalling the assembly line.5,13 One particularly intriguing feature of A domains is their ability to contain within them the catalytic region of other domains, typically interrupted between the a8 and a9 motifs,

atural products (NPs) are known to be a rich source of biologically active compounds with tremendous therapeutic relevance. Most medicines today have NP origins.1 While these NPs provide great therapeutic benefits by themselves, semisynthetic derivatives can improve their properties as seen in methylated somatostatin analogues, which have improved oral bioavailability compared to the parent compound.2 However, regiospecific modifications of NPs via traditional organic synthesis can be strenuous and challenging. Many NP scaffolds are not synthetically accessible due, in large part, to their high structural complexity, copious number of stereocenters, and annulated ring structures, and those that are synthetically accessible often require lengthy synthetic routes.3 Their direct chemical modification is also often hindered for these same reasons. Chemoenzymatic modifications of NPs offer an advantage over traditional synthesis for their derivatization since enzymatic chemistry tends to be regioand stereospecific. There is a significant need to utilize the potential of protein engineering to modify NPs and improve their biological activity. Nonribosomal peptides (NRPs) comprise a major class of NPs and are biosynthesized by mega-enzymes called nonribosomal peptide synthetases (NRPSs). NRPSs string together proteinogenic and nonproteinogenic amino acids to make a short peptide product.4 They are modular in nature and can therefore be subdivided into domains. These mega-enzymes function as an assembly line with each domain communicating with the next through communication-mediating linkers.5 Each NRPS module consists of three core domains: adenylation (A), condensation (C), and thiolation (T), which perform a particular function in the assembly line.4 The A domains are © XXXX American Chemical Society

Received: November 23, 2017

A

DOI: 10.1021/acssynbio.7b00426 ACS Synth. Biol. XXXX, XXX, XXX−XXX

Letter

ACS Synthetic Biology

Figure 1. (A) Structures of the natural products echinomycin, kutznerides, and thiocoraline. (B) Representation of the native (wild-type) and engineered enzymes investigated.

termed interrupted A domains.14 These interrupted A domains are bifunctional because they carry out both the adenylation of the amino acid substrate and the function of the inserted domain.15−18 By far, the most common interruption is by part of a methyltransferase (M) domain between the a8 and a9 motifs.14 Methylation in NPs is a common modification that can greatly enhance the biological activity of the molecules.2 The M domain portions of interrupted A domains, though incredibly intriguing, remain enigmatic, especially with regard to how interrupted A domains are generated and function cohesively. There are two types of M domains observed in interrupted A domains, those that carry out side chain (O- or S) methylation and those that carry out core (N-) methylation of the amino acid. There are two instances of methylation on the side chain of amino acid residues. The first example is the interrupted A domain KtzH(A4aMHA4bT4) of the kutznerides biosynthetic pathway (Figures 1A and S1), which carries out Omethylation of L-Ser side chain.16 The second example is the interrupted A domain, TioN(AaMNAb), which carries out Smethylation of L-Cys side chain.17 There are many examples of N-methylating M domains, particularly fascinating is TioS(A3aM3SA3b) from the thiocoraline biosynthetic pathway (Figures 1A and S1). TioS(A3aM3SA3b) has recently been characterized and shown to not only N-methylate L-Cys, but N,N-dimethylate it as well, a completely unprecedented finding.19 It was previously demonstrated that naturally interrupted A domains are susceptible to genetic manipulations when the MH portion of KtzH(A4aMHA4bT4) could be completely removed without affecting the adenylating activity of the enzyme.20 In the same study, KtzH(MH) was replaced with the MN portion of TioN(AaMNAb) to produce a functional interrupted A domain KtzH(A4aMNA4bT4) with an expanded substrate profile. Additionally, KtzH(A4aMNA4bT4) was able to load amino acid substrates onto its T domain, indicating that communication between domains was preserved. This is perhaps the strongest reason to investigate interrupted A domains as a potential way to structurally modify NRP products as this circumvents the issues seen with domains from different assembly lines being swapped or inserted into noncognate systems. Overall, interrupted A domains are fascinating bifunctional enzymes that warrant substantial investigation to elucidate their molecular mechanisms. Previous studies19,20 bring to light three

important questions surrounding interrupted A domains as potential biosynthetic tools: (i) Could a standard naturally occurring uninterrupted A domain be genetically engineered to contain a foreign domain, thus generating an artificial interrupted A domain? (ii) Would the newly transplanted noncognate M domains retain their methylating activity and specificity? (iii) Can two dissimilar M domains be used equally well? Herein, to answer these questions, we used the natural LSer activating A domain from the echinomycin biosynthetic pathway, Ecm6(A1T1),21 as a model M domain acceptor. Two very different M domains (Figure S2), the O-methylating interrupted A domain KtzH(A4aMHA4bT4)16 and the recently characterized N-methylating interrupted A domain TioS(A3aM3SA3b)19 were used as model side chain and core methylating M domain donors (Figure 1B) to interrupt Ecm6(A1T1) between its a8 and a9 region. These interruptions yielded Ecm6(A1aMHA1bT1) and Ecm6(A1aM3SA1bT1), respectively. We chose Ecm6(A1T1) as the M domain acceptor for two reasons: (i) because it activates L-Ser, which is the actual substrate of KtzH(MH), and (ii) because L-Ser is an isostere for L-Cys, which is the substrate for TioS(M3S). We anticipated that using an A domain which activates an amino acid that is the one or close to that usually methylated by the M domains inserted would be our best starting point. To evaluate the efficacy of these systems, the artificial interrupted A domains were evaluated in all aspects of A and M domain activity. This study serves as a proof-of-principle platform for the generation of bifunctional NRP (AMA) domains from monofunctional ones (A).

■

RESULTS AND DISCUSSION Cloning, Heterologous Co-overexpression, and Copurification of A Domains Used in This Study with Their MbtH-like Protein (MLP) Partners. To explore the possibility of generating functional artificially interrupted A domains we cloned four proteins: the wild-type (wt) natural monofunctional Ecm6(A1T1), the artificially interrupted Ecm6(A1aMHA1bT1) and Ecm6(A1aM3SA1bT1), and the TioS(A3aM3SA3b)D599A mutant lacking methylating activity. We utilized a standard cloning method for the native A domain Ecm6(A1T1) and single overlap extension to generate the point mutant TioS(A3aM3SA3b)D599A. To produce the artificially interrupted A domains Ecm6(A1aMHA1bT1) and Ecm6B

DOI: 10.1021/acssynbio.7b00426 ACS Synth. Biol. XXXX, XXX, XXX−XXX

Letter

ACS Synthetic Biology (A1aM3SA1bT1), we employed a versatile strategy for facile assembly of the interrupted constructs (Figure S3). We heterologously co-overexpressed and copurified these enzymes in E. coli BL21 (DE3) with their native MLP partners (Ecm8 and TioT for the Ecm6 and TioS proteins, respectively) as MLPs have previously been reported to be necessary for soluble expression and/or activity of some A domains,22,23 including those from the echinomycin,24 kutznerides,16 and thiocoraline,17,25 biosynthetic pathways. Characterization of the Adenylating Activity of Ecm6(A1T1), Ecm6(A1aMHA1bT1), and Ecm6(A1aM3SA1bT1). A major obstacle in engineering biosynthetic enzymes is the diminished activity observed once the protein has undergone genetic modifications, which often requires additional modifications to restore activity.26 In order to develop usable engineered interrupted A domains for the purpose of NP modification, several criteria must be met. The engineered A domain should (i) retain the same or similar substrate specificity as that of the wt enzyme, as changing it can sometimes affect downstream reactions, (ii) maintain a similar level of adenylating activity to that of the wt enzyme, and (iii) contain an insertion that does not interfere with the function or activity of the downstream T domain. To evaluate the ability of the engineered Ecm6(A1aMHA1bT1) and Ecm6(A1aM3SA1bT1) to meet these requirements, we first determined the substrate specificity profile of the wt Ecm6(A1T1) and that of the engineered Ecm6(A1aMHA1bT1) and Ecm6(A1aM3SA1bT1) using an ATP-[32P]PPi exchange assay. We evaluated 20 natural Lamino acids and two methylated Ser derivatives (N-Me-L-Ser and O-Me-D,L-Ser) as potential substrates of these enzymes (Figure 2). The substrate profiles revealed that Ecm6-

Table 1. Steady-State Kinetic Parameters for AMP Derivatization of L-Ser by Ecm6(A1T1), Ecm6(A1aMHA1bT1), and Ecm6(A1aM3SA1bT1) protein

Km (mM)

kcat (min−1)

kcat/Km (mM−1 min−1)

Ecm6(A1T1) Ecm6(A1aMHA1bT1) Ecm6(A1aM3SA1bT1)

1.5 ± 0.4 5.3 ± 0.8 6.6 ± 0.6

5.8 ± 0.6 2.29 ± 0.02 25 ± 1

3.9 ± 1.2 0.43 ± 0.07 3.7 ± 0.4

substrate affinity (Km) of the enzymes, generally increasing it by about 4-fold. However, the KtzH(MH) and TioS(M3S) insertions changed the turnover rates (kcat) in an opposite manner. When compared to Ecm6(A1T1), Ecm6(A1aMHA1bT1) showed a decrease in kcat by 2.5-fold in contrast to Ecm6(A1aM3SA1bT1), which showed an increase in kcat by 4fold. The changes in Km and kcat were reflected in the overall catalytic efficiency (kcat/Km). When compared to the wt Ecm6(A1T1), Ecm6(A1aMHA1bT1) displayed a decrease in kcat/Km by 9-fold, and Ecm6(A1aM3SA1bT1) had essentially no change in the kcat/Km, indicating that it is just as catalytically efficient as the wt enzyme. These kinetic data fulfilled criterion (ii), especially with regard to Ecm6(A1aM3SA1bT1). Finally, to verify that the insertion of an M domain does not interfere with the function of the downstream T domain, we assessed by TCA precipitation assay using [3H]acetyl-CoA and the promiscuous PPTase, Sfp,27,28 the ability of the T domains of all three proteins to be converted from their apo to holo state (Figure S6). These experiments revealed that the T domains of both Ecm6(A1aMHA1bT1) and Ecm6(A1aM3SA1bT1) were able to undergo holo conversion. Additionally, since the purpose of the A domain activating the amino acid is to allow for the substrate to be loaded onto the holo T domain, we monitored the activation and subsequent loading of [3H]L-Ser by TCA precipitation assay and found that the T domains of both interrupted enzymes were able to be loaded (Figure S7). All the above data combined indicate that the normal functions of the A and T domains and their interaction with each other remained intact while harboring either noncognate M domain. These data fulfilled the last criterion (iii), indicating that all three criteria required for developing the engineered interrupted A domains were met. Evaluation of the Methylating Activity of Ecm6(A1aMHA1bT1) and Ecm6(A1aM3SA1bT1). Having established that the normal operations of the A and T domains remain functional after insertion of a noncognate M domain, we next sought to explore the activity of the inserted M domains. We monitored the methylation events via time course TCA precipitation assays with L-Ser and [methyl-3H]S-adenosyl-Lmethionine (SAM) (Figure 3). Since it was previously shown that the substrate needs to be tethered to the T domain in order to be methylated,19 we first prepared a reaction mixture that generated loaded L-Ser-S-Ecm6(A1aMHA1bT1) or L-Ser-SEcm6(A1aM3SA1bT1). Once loading of the L-Ser was complete, the methylation reaction was initiated by addition of [methyl-3H]SAM and monitored over time. We found that the engineered interrupted A domains were able to carry out methylation of L-Ser. Collectively, these experiments revealed that Ecm6(A1aMHA1bT1) and Ecm6(A1aM3SA1bT1) are fully functional as they can perform adenylation of an amino acid followed by its loading onto T domains, and subsequent methylation. Elucidation of M Domain Methylation Specificity. After establishing that the engineered interrupted A domains,

Figure 2. Substrate profiles of wild-type (wt) Ecm6(A1T1) and engineered interrupted A domains Ecm6(A1aMHA1bT1) and Ecm6(A1aM3SA1bT1) determined by ATP-[32P]PPi exchange assays at a 2-h end point.

(A1aMHA1bT1) and Ecm6(A1aM3SA1bT1) displayed specificity for L-Ser, which was also observed with Ecm6(A1T1), satisfying criterion (i). Next, to investigate the level of adenylating activity of Ecm6(A1T1), Ecm6(A1aMHA1bT1), and Ecm6(A1aM3SA1bT1), we determined their Michaelis−Menten steady-state kinetic parameters for adenylation of L-Ser (Table 1 and Figure S5). Our results revealed that the insertion of both KtzH(MH) and TioS(M3S) into Ecm6(A1T1) had a similar effect on the C

DOI: 10.1021/acssynbio.7b00426 ACS Synth. Biol. XXXX, XXX, XXX−XXX

Letter

ACS Synthetic Biology

control Ecm6(A1T1) (Figure 4A). Additionally, TioN(AaMNAb) has been shown to adenylate, but not methylate N,S-diMe-L-

Figure 3. Methylation of L-Ser by (A) Ecm6(A1T1) and Ecm6(A1aMHA1bT1), and (B) Ecm6(A1T1) and Ecm6(A1aM3SA1bT1) with [methyl-3H]SAM as the methylating agent, measured by TCA precipitation assays.

Ecm6(A1aMHA1bT1) and Ecm6(A1aM3SA1bT1), were capable of methylating L-Ser, we next scrutinized the methylation events in detail. It is important to note that the M domains KtzH(MH) and TioS(M3S) used for insertion (i) share only 43% amino acid sequence identity (Figure S2) and (ii) are proposed to catalyze side chain O-methylation and core N-methylation, respectively.16,19 Since methylation occurs once the substrate is loaded onto the T domain,19 we were limited to amino acids that could be activated and loaded onto the T domains of Ecm6(A1aMHA1bT1) and Ecm6(A1aM3SA1bT1). As established during the substrate profile experiment (Figure 2), O-Me-D,LSer or N-Me-L-Ser are not substrates for Ecm6(A1aMHA1bT1) and Ecm6(A1aM3SA1bT1). To circumvent this, we used an alternative A domain to activate L-Cys, a suitable isostere of LSer and the natural substrate of TioS(M3S). For this, we generated a point mutant of TioS(A3aM3SA3b), which has a known substrate profile,19 that would activate, but not methylate two of the substrates needed. To inactivate the methylating activity of the wt TioS(A3aM3SA3b), we identified a key amino acid residue in the M domain of TioS(A3aM3SA3b) based on a homology alignment of TioS(A3aM3SA3b) with known methyltransferases. The key residue identified, Asp599, was then mutated to Ala to create TioS(A3aM3SA3b)D599A. We then verified that TioS(A3aM3SA3b)D599A could adenylate NMe-L-Cys and S-Me-L-Cys (Figure S4). We also confirmed the lack of methylating activity of this mutant by observing no increase in the methylation assay for the nonmethylating wt

Figure 4. Methylation with [methyl-3H]SAM as the methylating agent, measured by TCA precipitation assays of (A) L-Cys by Ecm6(A1T1), Ecm6(A1aMHA1bT1), and Ecm6(A1aM3SA1bT1), (B) L-Cys, N-Me-LCys, and S-Me-L-Cys by Ecm6(A1aM3SA1bT1), and (C) N,S-diMe-LCys by Ecm6(A1T1) and Ecm6(A1aM3SA1bT1).

Cys.19 By using N-Me-L-Cys and S-Me-L-Cys as substrates for the M domain, we are able to control the methylation sites available on the amino acid. With N-Me-L-Cys, only Smethylation or a second methylation on the N are available, and with S-Me-L-Cys, only N-methylation is possible. When using N,S-diMe-L-Cys as a substrate, the only place for methylation to occur is a second methylation event on the N. D

DOI: 10.1021/acssynbio.7b00426 ACS Synth. Biol. XXXX, XXX, XXX−XXX

Letter

ACS Synthetic Biology

modifying NRPs. Such work is currently in progress in our laboratory.

First, we evaluated if L-Cys could be methylated by Ecm6(A1aMHA1bT1) and Ecm6(A1aM3SA1bT1) (Figure 4A). This assay revealed that while Ecm6(A1aMHA1bT1) could not methylate L-Cys, Ecm6(A1aM3SA1bT1) could. These data were not entirely unexpected as KtzH(MH) has been proposed to carry out O-methylation, and it is plausible that sulfur is too large to properly fit in the active site of KtzH(MH). Therefore, these data suggest KtzH(MH) specifically catalyzes Omethylation. The natural substrate of TioS(M3S), however, is L-Cys, which is consistent with these data showing that Ecm6(A1aM3SA1bT1) can methylate L-Cys. To further characterize Ecm6(A1aM3SA1bT1), we used TioS(A3aM3SA3b)D599A to activate and load L-Cys, N-Me-LCys, or S-Me-L-Cys onto the T domain of Ecm6(A1aM3SA1bT1) and monitored methylation over time (Figure 4B). We detected methylation occurring for all three substrates indicating that Ecm6(A1aM3SA1bT1) likely methylates the core N of L-Cys. Furthermore, these methylation data suggest that, like the wt M domain donor TioS(A3aM3SA3b), the engineered Ecm6(A1aM3SA1bT1) can carry out dimethylation on the N of LCys. To confirm this, we used TioN(AaMNAb) to activate and load N,S-diMe- L -Cys onto the T domain of Ecm6(A1aM3SA1bT1) and monitored the methylation event (Figure 4C). As observed with TioS(A 3 a M 3 S A 3 b ), 1 9 Ecm6(A1aM3SA1bT1) could methylate N,S-diMe-L-Cys, supporting a dimethylation event on the N. Together, these data support that the transplanted M domains do in fact retain the methylation pattern and specificity observed for the original enzymes from which they came. Additionally, these data suggest that we can site-specifically methylate a particular amino acid by interrupting an A domain with an M domain that is known to carry out that type of methylation.

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acssynbio.7b00426. Procedures for all experiments performed; Figures S1− S7; Table S1 (PDF)

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. ORCID

Sylvie Garneau-Tsodikova: 0000-0002-7961-5555 Author Contributions

T.A.L. and S.G.-T. designed the study. T.A.L. performed all biochemical experiments with the exception of elucidation of M domain methylation specificity, which was done by S.M. T.A.L. and S.G.-T. wrote the paper and Supporting Information. S.M. helped with writing of the Supporting Information and proofreading of the manuscript. All the figures were generated by T.A.L. and S.G.-T. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS We thank Dr. Atefeh Garzan for providing the N-Me-L-Cys and N,S-di-Me-L-Cys, for which the synthesis was previously published in ref 19. This project was supported by startup funds from the College of Pharmacy at the University of Kentucky (to S.G.-T.) and by a National Science Foundation CAREER Award MCB-11494278 (to S.G.-T).

■

CONCLUSION In summary, we have created two completely novel types of fully functional engineered interrupted A domains by inserting foreign (side chain and core methylating) M domains into a naturally occurring monofunctinal A domain to emulate natural interrupted A domains. We established a proof-of-concept that (i) artificial interrupted A domains can be engineered to contain noncognate M domains, thus achieving bifunctionality, (ii) both the A and M domains retain the same adenylating and methylating functions as their original sources, and (iii) the identity of the M domain dictates the location of methylation (side chain or core). As exhilarating as these data are, this work provokes many new questions to be addressed in the future. At the forefront of the questions in line stands: can other auxiliary domains be used to interrupt A domains in the same fashion as done here? Can M domains that have been modified to accept non-natural SAM analogues be utilized to afford a chemical handle for further synthetic modifications? Can this be done with other non-NRPS enzymes? Additionally, this work offers an alternative approach to introducing site specific modifications to NRPS systems, which traditionally has utilized domain swaps followed by directed evolution. Since this is the first report of engineering these types of bifunctional enzymes from monofunctional ones, we employed a simple model system to get at the root question of this research; is this even possible? Future work will aim to explore other A domains and other auxiliary domains to determine the overall versatility of this approach. Additionally, this work provides the foundation for future long-term work to express these and other engineered interrupted A domains in vivo with the aim of site-specifically

■

REFERENCES

(1) Newman, D. J., and Cragg, G. M. (2016) Natural products as sources of new drugs from 1981 to 2014. J. Nat. Prod. 79, 629−661. (2) Biron, E., Chatterjee, J., Ovadia, O., Langenegger, D., Brueggen, J., Hoyer, D., Schmid, H. A., Jelinek, R., Gilon, C., Hoffman, A., and Kessler, H. (2008) Improving oral bioavailability of peptides by multiple N-methylation: somatostatin analogues. Angew. Chem., Int. Ed. 47, 2595−2599. (3) Nicolaou, K. C., Yang, Z., Liu, J. J., Ueno, H., Nantermet, P. G., Guy, R. K., Claiborne, C. F., Renaud, J., Couladouros, E. A., Paulvannan, K., et al. (1994) Total synthesis of taxol. Nature 367, 630−634. (4) Schwarzer, D., Finking, R., and Marahiel, M. A. (2003) Nonribosomal peptides: from genes to products. Nat. Prod. Rep. 20, 275−287. (5) Hahn, M., and Stachelhaus, T. (2004) Selective interaction between nonribosomal peptide synthetases is facilitated by short communication-mediating domains. Proc. Natl. Acad. Sci. U. S. A. 101, 15585−15590. (6) Walsh, C. T., Gehring, A. M., Weinreb, P. H., Quadri, L. E., and Flugel, R. S. (1997) Post-translational modification of polyketide and nonribosomal peptide synthases. Curr. Opin. Chem. Biol. 1, 309−315. (7) Konz, D., and Marahiel, M. A. (1999) How do peptide synthetases generate structural diversity? Chem. Biol. 6, R39−48. (8) Conti, E., Stachelhaus, T., Marahiel, M. A., and Brick, P. (1997) Structural basis for the activation of phenylalanine in the nonribosomal biosynthesis of gramicidin S. EMBO J. 16, 4174−4183. (9) Winn, M., Fyans, J. K., Zhuo, Y., and Micklefield, J. (2016) Recent advances in engineering nonribosomal peptide assembly lines. Nat. Prod. Rep. 33, 317−347.

E

DOI: 10.1021/acssynbio.7b00426 ACS Synth. Biol. XXXX, XXX, XXX−XXX

Letter

ACS Synthetic Biology (10) Villiers, B., and Hollfelder, F. (2011) Directed evolution of a gatekeeper domain in nonribosomal peptide synthesis. Chem. Biol. 18, 1290−1299. (11) Thirlway, J., Lewis, R., Nunns, L., Al Nakeeb, M., Styles, M., Struck, A. W., Smith, C. P., and Micklefield, J. (2012) Introduction of a non-natural amino acid into a nonribosomal peptide antibiotic by modification of adenylation domain specificity. Angew. Chem., Int. Ed. 51, 7181−7184. (12) Kries, H., Niquille, D. L., and Hilvert, D. (2015) A subdomain swap strategy for reengineering nonribosomal peptides. Chem. Biol. 22, 640−648. (13) Liu, H., Gao, L., Han, J., Ma, Z., Lu, Z., Dai, C., Zhang, C., and Bie, X. (2016) Biocombinatorial synthesis of novel lipopeptides by COM domain-mediated reprogramming of the plipastatin NRPS complex. Front. Microbiol. 7, 1801. (14) Labby, K. J., Watsula, S. G., and Garneau-Tsodikova, S. (2015) Interrupted adenylation domains: unique bifunctional enzymes involved in nonribosomal peptide biosynthesis. Nat. Prod. Rep. 32, 641−653. (15) Hornbogen, T., Riechers, S. P., Prinz, B., Schultchen, J., Lang, C., Schmidt, S., Mugge, C., Turkanovic, S., Sussmuth, R. D., Tauberger, E., and Zocher, R. (2007) Functional characterization of the recombinant N-methyltransferase domain from the multienzyme enniatin synthetase. ChemBioChem 8, 1048−1054. (16) Zolova, O. E., and Garneau-Tsodikova, S. (2014) KtzJdependent serine activation and O-methylation by KtzH for kutznerides biosynthesis. J. Antibiot. 67, 59−64. (17) Al-Mestarihi, A. H., Villamizar, G., Fernandez, J., Zolova, O. E., Lombo, F., and Garneau-Tsodikova, S. (2014) Adenylation and Smethylation of cysteine by the bifunctional enzyme TioN in thiocoraline biosynthesis. J. Am. Chem. Soc. 136, 17350−17354. (18) Magarvey, N. A., Ehling-Schulz, M., and Walsh, C. T. (2006) Characterization of the cereulide NRPS alpha-hydroxy acid specifying modules: activation of alpha-keto acids and chiral reduction on the assembly line. J. Am. Chem. Soc. 128, 10698−10699. (19) Mori, S., Garzan, A., Tsodikov, O. V., and Garneau-Tsodikova, S. (2017) Deciphering Nature’s intricate way of N,S-dimethylating Lcysteine: Sequential action of two bifunctional adenylation domains. Biochemistry 56, 6087−6097. (20) Shrestha, S. K., and Garneau-Tsodikova, S. (2016) Expanding substrate promiscuity by engineering a novel adenylating-methylating NRPS bifunctional enzyme. ChemBioChem 17, 1328−1332. (21) Watanabe, K., Hotta, K., Praseuth, A. P., Koketsu, K., Migita, A., Boddy, C. N., Wang, C. C., Oguri, H., and Oikawa, H. (2006) Total biosynthesis of antitumor nonribosomal peptides in Escherichia coli. Nat. Chem. Biol. 2, 423−428. (22) Zhang, W., Heemstra, J. R., Jr., Walsh, C. T., and Imker, H. J. (2010) Activation of the pacidamycin PacL adenylation domain by MbtH-like proteins. Biochemistry 49, 9946−9947. (23) Felnagle, E. A., Barkei, J. J., Park, H., Podevels, A. M., McMahon, M. D., Drott, D. W., and Thomas, M. G. (2010) MbtH-like proteins as integral components of bacterial nonribosomal peptide synthetases. Biochemistry 49, 8815−8817. (24) Zhang, C., Kong, L., Liu, Q., Lei, X., Zhu, T., Yin, J., Lin, B., Deng, Z., and You, D. (2013) In vitro characterization of echinomycin biosynthesis: formation and hydroxylation of L-tryptophanyl-S-enzyme and oxidation of (2S,3S) beta-hydroxytryptophan. PLoS One 8, e56772. (25) Zolova, O. E., and Garneau-Tsodikova, S. (2012) Importance of the MbtH-like protein TioT for production and activation of the thiocoraline adenylation domain of TioK. MedChemComm 3, 950− 955. (26) Fischbach, M. A., Lai, J. R., Roche, E. D., Walsh, C. T., and Liu, D. R. (2007) Directed evolution can rapidly improve the activity of chimeric assembly-line enzymes. Proc. Natl. Acad. Sci. U. S. A. 104, 11951−19956. (27) Quadri, L. E., Weinreb, P. H., Lei, M., Nakano, M. M., Zuber, P., and Walsh, C. T. (1998) Characterization of Sfp, a Bacillus subtilis

phosphopantetheinyl transferase for peptidyl carrier protein domains in peptide synthetases. Biochemistry 37, 1585−1595. (28) Yang, G., Zhang, Y., Lee, N. K., Cozad, M. A., Kearney, S. E., Luesch, H., and Ding, Y. (2017) Cyanobacterial Sfp-type phosphopantetheinyl transferases functionalize carrier proteins of diverse biosynthetic pathways. Sci. Rep. 7, 11888.

F

DOI: 10.1021/acssynbio.7b00426 ACS Synth. Biol. XXXX, XXX, XXX−XXX