The specificity of intermodular recognition in a prototypical nonribosomal peptide synthetase depends on an adaptor domain

Abstract

In the quest for new bioactive substances, nonribosomal peptide synthetases (NRPS) provide biodiversity by synthesizing nonproteinaceous peptides with high cellular activity. NRPS machinery consists of multiple modules, each catalyzing a unique series of chemical reactions. Incomplete understanding of the biophysical principles orchestrating these reaction arrays limits the exploitation of NRPSs in synthetic biology. Here, we use nuclear magnetic resonance (NMR) spectroscopy and mass spectrometry to solve the conundrum of how intermodular recognition is coupled with loaded carrier protein specificity in the tomaymycin NRPS. We discover an adaptor domain that directly recruits the loaded carrier protein from the initiation module to the elongation module and reveal its mechanism of action. The adaptor domain of the type found here has specificity rules that could potentially be exploited in the design of engineered NRPS machinery.

Fig. 1.. Condensation reaction of the tomaymycin NRPS.

(A) Schematic models of prototypical initiation and elongation modules performing adenylation (left) and condensation (right) reactions. The initiation module comprises the adenylation domain A and the peptidyl carrier domain PCP. Its substrate is represented by the black star. The elongation module comprises the condensation domain C, the adenylation domain A, and the peptidyl carrier domain PCP. Its substrate is represented by the black hexagon. (B) Architecture of the tomaymycin NRPS: initiation module (TomA) and elongation/termination module TomB, containing the reductase domain BR. The two natural substrates and the product tomaymycin are displayed underneath. aa, amino acid. (C) Substrates used in this study (shown conjugated to the terminal thiol of ppant) and the products obtained after the BC-catalyzed condensation reaction and spontaneous ring closure. (D and E) LC-MS analysis of the LC peak containing the product of the condensation reaction and subsequent spontaneous ring closure, as shown in (C). The expected product mass on the M⁺ channel is 229 Da. The horizontal and vertical axes represent the mass/charge ratio (m/z) and relative abundance, respectively. The peak at 251 Da is TCEP. (D) and (E) show the reactions conducted with excised domains and full length proteins, respectively.

Fig. 2.. Structures of excised APCP and BN-BC domains.

(A) Crystallographic structure of BN-BC at 1.7-Å resolution (N-form). The N- and the C-lobes are shown in light and dark green, respectively. The two cross-over elements—the floor loop and latch loop—are shown in red and beige, respectively. Also highlighted in red is the junction between the N- and C-lobes. The first 79 amino acids do not yield any electron density. (B) Structure of BN-BC showing the sites on opposite faces of the protein where the donor and acceptor PCPs are expected to bind. (C) NMR structure of APCP_ppant (amino acids 5 to 81). The disordered ppant arm is omitted for clarity. (D) NMR structure of substrate-loaded APCP_ppant (APCP_load; amino acids 5 to 81) showing both the ppant arm and the substrate (in stick representation). (E) NMR structure of BN (amino acids 4 to 77), which closely resembles the PCP fold.

Fig. 3.. BC exhibits conformational heterogeneity.

(A and B) Excerpts of the ¹H,¹³C-methyl HMQC spectrum of 50 μM ¹H,¹³C-ILV-methyl–labeled ²H-BN-BC with peak assignments. Asterisks indicate 35 unassigned peaks. ppm, parts per million. (C and D) Expansions of regions from the ¹H,¹³C-methyl HMQC spectrum showing methyl groups that adopt two conformations in both isolated BN-BC (left) and after addition of 12 equiv of APCP_load (right; BN-BC at a concentration of 100 μM in both). Purple and red labels indicate the major and minor conformations, respectively. The population of the minor conformation increases in the presence of APCP_load. (E) For some methyl groups, the peak corresponding to the minor conformation increases in intensity and shifts to a new position in the presence of APCP_load (light blue). (F and G) Some of the peaks showing little or no evidence of a second conformation in isolated BN-BC split into two peaks upon addition of APCP_load, corresponding to the APCP-bound (light blue) and unbound (purple) forms. (H) Peaks belonging to the BN domain in the BN-BC construct show a bimodal behavior upon addition of 1.8 to 12 equiv of APCP_load to 50 μM BN-BC (spectra are color-coded according to the number of APCP_load equivalents).

Fig. 4.. Structures of the encounter and final BN-BC–APCP_load complexes.

(A) Structure of the encounter complex formed between the BN domain of BN-BC (in orange) and APCP_load (in blue). The ppant arm is omitted for clarity, but the red dots indicate the position of the modified serine. Hydrogen bonds and hydrophobic contacts are shown in the upper and bottom views, respectively. (B) Overview of the structure of the final BN-BC–APCP_load complex (BC, green; BN, orange; APCP_load, blue). (C to E) Close-up views of the atomic interactions between APCP-α3 (C), APCP-F61 (D), and APCP-α2 (E) with BC. (F and G) Close-up views of the atomic interactions between BN and APCP-α1 (F) and APCP-α2 as well as BC (G).

Fig. 5.. Schematic model of APCP recognition by BN-BC.

BN encompasses a folded PCP-like domain, which recruits APCP from solution through formation of an encounter complex. Because of the flexibility provided by the disordered region corresponding to amino acids 78 to 90, the APCP bound to BN can scan the surface of BC. Once APCP finds the donor binding site, an array of specific (but individually weak) interactions commits APCP to this binding site. BN readjusts its position to optimize the interaction interfaces and yield the final complex.