The Biochemistry and Structural Biology of Cyanobactin Pathways: Enabling Combinatorial Biosynthesis

Abstract

Cyanobactin biosynthetic enzymes have exceptional versatility in the synthesis of natural and unnatural products. Cyanobactins are ribosomally synthesized and posttranslationally modified peptides synthesized by multistep pathways involving a broad suite of enzymes, including heterocyclases/cyclodehydratases, macrocyclases, proteases, prenyltransferases, methyltransferases, and others. Here, we describe the enzymology and structural biology of cyanobactin biosynthetic enzymes, aiming at the twin goals of understanding biochemical mechanisms and biosynthetic plasticity. We highlight how this common suite of enzymes may be utilized to generate a large array or structurally and chemically diverse compounds.

Keywords: Combinatorial biosynthesis; Cyanobactins; Cyclic peptide; Natural products; RiPP biosynthesis; Synthetic biology.

Figure 1.. Representative cyanobactins.

Posttranslational modifications are indicated by colors, tied to enzymes that catalyze each transformation. * indicates hypothetical reaction, all others are experimentally defined.

Figure 2.. Canonical cyanobactin biosynthetic pathway.

Top: The pat pathway to patellamides, with an expansion showing the amino acid sequence of PatE1, the substrate for biosynthetic enzymes. The precursor peptide substrate may be demarcated as described in the Text Box. The order of the modification occurs such that heterocyclization occurs first, followed by proteolysis and macrocyclization. Late stage tailoring may further elaborate the modified peptides.

Figure 3.. Recognition sequences and precursor peptides.

Shown are example sequences of cyanobactin precursors, with color-coded RSs tied to modifying enzymes. The domain architecture of each enzyme is laid out adjacent to the enzyme name. Prenyltransferase does not require a recognition sequence. Many cyanobactin precursor peptides contain multiple core sequences that are processed into different products.

Figure 4.. Heterocyclase reactions.

A. TruD and PatD bind to RSI and catalyze heterocycle formation. TruD, ~100% identical to PatD in the RRE and E1-like domain, is 77% identical to PatD in YcaO domain. PatD synthesizes thiazoline and oxazoline, while TruD is a thiazoline specialist. As defined by Walsh, thiazoline is energetically more favorable (Belshaw et al., 1998). B. TruD converts ATP to AMP and PPi, whereas some other heterocyclases such as those involved in linear azol(in)e-containing peptides (LAP) or thiopeptide biosynthesis yield ADP and Pi as the products.

Figure 5.. Sequence similarity analysis of heterocyclases.

A. All heterocyclase domain-containing proteins. B. Expansion of cyanobacterial heterocyclase group and close relatives. The cyanobactin group is found in 4b of this branch.

Figure 6.. Structures of the cyanobactin heterocyclase.

Crystal structures of LynD (PDB code 4V1T) showing RSI from the PatE precursor peptide (in green) bound to the RRE (in blue) and ATP in the YcaO active site (right, in pink).

Figure 7.. Macrocyclization by PatA and PatG.

PatA frees an N-terminus, which is then available for macrocyclization by PatG.

Figure 8.. Sequence similarity diagram of bacterial S8A proteases.

A. Proteases from a broad swath of bacteria. B. Expansion of red box from panel A. Showing the close relationship between PatA- and PatG-like proteases. BGC: biosynthetic gene cluster.

Figure 9.. Macrocyclase rules.

Macrocyclase PatG is broadly tolerant of substrates and is capable of generating circular, or sometimes linear products. The above rules are summarized from hundreds of substrates that have been analyzed. Recognition “rules” are useful in synthetic biology and in understanding enzyme mechanism, but there are always exceptions. For example, glycine is disfavored in position P2, but there are examples in which it is successfully used. In rule 3, NR = no reaction, L = linear product, C:L = both linear and cyclic product, and C = cyclic product.

Figure 10.. Crystal structures of PatA and PatG protease domains.

The proteins have very similar architectures except for a cap in PatG, shown here interacting with an artificial substrate (shown as yellow sticks). The panel on the right shows a close-up view of the PatG active site, in the vicinity of the scissile bond.

Figure 11.. Mechanism of macrocyclase.

The cap (wavy red) interacts with RSIII (red) to hold the substrate in place. A catalytic triad typical of S8A proteases forms a covalent adduct with P1 at Ser783; this is displaced by an internal nucleophilic nitrogen to afford a macrocycle, while in most proteases hydrolysis would occur instead. The participation of individual residues in catalysis (other than Ser783) remains speculative.

Figure 12.. Sequence similarity diagram of PatF-like prenyltransferases.

Note that sequence clusters shown in Network diagram (left) so far fairly accurately predict the chemistry catalyzed by the enzymes within each cluster (right).

Figure 13.. Prenylation by cyanobactin enzymes.

Shown is a single prenylation of the trunkamide precursor peptide by TruF1.

Figure 14.. Structure of PagF bound to substrate.

While most ABBA prenyltransferases contain a hydrophobic active site that precludes solvent, the structure of PagF reveals a solvent exposed cavity. Binding of the peptide substrate is necessary to form a solvent excluded cavity where productive prenyl transfer can occur.

Figure 15.

Aeruginosamide biosynthesis via the age pathway.

Figure 16.

General steps for in vitro construction of artificial cyanobactins.