Even though the ribosome is the molecular machinery mostly used by cells to produce complex polypeptides, bacteria [1, 2] and fungi [3, 4] have an additional pathway that synthesizes peptides via non-ribosomal peptide synthetases (NRPSs). The short peptides assembled by NRPS, termed non-ribosomal peptides (NRPs), comprise a wide range of secondary metabolites, including commonly used antibiotics as well as metabolic and detoxifying enzymes [5, 6]. NRPs range in size from 2-48 amino acid residues [6, 7, 8] and are of remarkable structural variability. In contrast to the ribosomal synthesis pathway which is mainly restricted to the 21 proteinogenic amino acids [9, 10], NRPSs accept more than 500 different monomers as substrates, including non-proteinogenic, N-methylated and D-amino acids [11]. Although the functionality and structure of their synthesized products are of high diversity, NRPS are characterized by a common structural theme: their composition of distinct modular sections (reviewed in [12, 13]). Therefore, the study of NRPSs allows for the exploration of one of the fundamental principles of synthetic biology: modularity.

NRPS Are of Modular Structure

NRPS Are Organized in Biosynthetic Gene Clusters

The genes encoding for NRPS are organized in a large gene cluster [14], so called assembly lines. These contain i) the genes coding for the non-ribosomal peptide synthetases, which can reach a size of up to 1.5 MDa e. g. cyclosporine [15], ii) genes necessary for the biosynthesis of the monomers to be incorporated into the NRP, and iii) tailoring enzymes introducing further modifications in the peptide [16]. Often, the gene clusters code for more than just one peptide synthetase [17]. These co-encoded enzymes are commonly connected by communication domains in order to keep the structure of the assembly line [17]. Thus, the cloning of complete NRPS pathways is facilitated by the clustered structure of the NRPS encoding genes.

One Module - One Amino Acid

Each NRPS is organized in so called modules, where a single module serves as a catalytic unit responsible for the incorporating of one amino acid in the growing peptide chain. The order of the catalytic units defines the sequence and the chemical characteristics of the resulting NRP where the number and order of the units corresponds precisely to the number and order of the amino acids in the resulting peptide [15]. The modules itself are of comprised of smaller subunits termed domains. Each domain fulfills a distinct function in the assembly of the growing peptide chain [13]. The two levels of modularity - modules and domains are shown in Fig. 1.

Principle of Non-Ribosomal Peptide Synthesis

For the synthesis of NRPs at least three different domains are essential to carry out the selection of the substrate, its activation and the peptide bond formation: Adenylation, thiolation and condensation domains [13].

Activation of the Thiolation Domains by Phosphopantetheinyl-Tranferases

NRP synthetases contain thiolation domains (T domains) in every module, where the corresponding amino acid is covalently bound to the enzyme via a thioester bond. The essential amino acid residue of the T domains is a serine which is posttranslationally modified to carry the sulfhydryl group required for the thioester bond formation. This modification is carried out by separate proteins - the phosphopantetheinyltranferases (PPTases), which use coenzyme A as cofactor [18]. The reaction is shown in Fig. 2a.

Activation of Monomers by Adenylation Domains

The attachment of the amino acid to the already activated T domain is carried out by the adenylation domain (A domain) which is highly substrate specific for a single monomer. In a first step, the monomer to be incorporated is activated with ATP. Secondly, the thioester bond between the phosphopantetheinyl residue and the monomer's carboxylic acid residue is formed [19]. The reaction is depicted in Fig. 2b.

Formation of Peptide Bonds by Condensation Domains

When two neighboring monomers are activated, the condensation domain (C) catalyzes formation of the peptide bond, which is shown in Fig. 2c. The C domain is selective for the acceptor amino acid and thus the C and A domain of a given module always display the same substrate specificity [20].

Termination of NRP Assembly by Thioesterases

In order to terminate the NRP assembly, a thioesterase (TE) domain is located at the end of most NRPSs. TE domains act in two steps. They first bind the peptide to their own conserved serine residue and in a second step this bond is attacked by a nucleophile [21]. The functionalities of the TE domains can roughly be classified into i) product cleavage [22] and ii) macro cyclization [23]. Those TE domains belonging to the former group cleave the thioester bond between the product and the enzyme complex by allowing for a nucleophilic attack with water. TE domains of the second group introduce a macro cycle into the product by an intramolecular nucleophilic attack [21]. This macro cycle can either connect the two peptide termini (yielding a closed ring structure such as tyrocidine [23]) or introduce any other peptide bond based cycle (yielding partially circularized NRPs for instance bacitracin [24]). In addition to those integral TE domains, independent but cluster-encoded thioesterases are known to function as a rescue protein for stalled NRPS [25].

Monomer Modifications

NRPS can incorporate more than 500 different monomers. These monomers include proteinogenic and non-proteinogenic amino acids as well as arylic acids. The latter lack the amine residue and are thus only suitable for chain initiations [11].

Incorporation of D-Amino Acids Is Facilitated by Epimerisation Domains

The most basic derivation is the introduction of D-amino acids into NRPs. Their incorporation is achieved by epimerisation (E) domains located between the T domain of the donor and the C domain of the acceptor module (see Fig. 3). It racemises the donor amino acid and the correct stereo-conformation is selected by the acceptor C domain. Thus the C domains can be classified as either C_D or C_L specific [26].

Methylation of Monomers by N-Methyltransferases

Another basic modification of monomers is N-methylation carried out by N-methyltransferases (NM). They are located between the A and the T domain of a module and transfer CH₃ from S-adenosylmethionine to the amino group of the module's monomer after it has been activated by the A domain [27].

Heterocyclisation of NRPs by Cyclisation Domains

Besides the macro-cyclisation performed by TEdomains, cyclisation (Cy) domains can introduce additional amide bonds between neighboring amino acids within the assembly line. In order to form a heterocycle, serine, threonine or cystein are required as acceptor amino acid [28].

As the borders of the C or A domains can be hard to define, Cy domains and E domains are often displayed as combined C/Cy or C/E-domains. Besides these most common modifications many others can be found within the NRPS cluster or as tailoring enzymes, for instance O-methylations, halogenations, glycosylations or oxidations [13] - there is a lot more to explore!