Total Synthesis and Structure Assignment of the Relacidine Lipopeptide Antibiotics and Preparation of Analogues with Enhanced Stability

The unabated rise of antibiotic resistance has raised the specter of a post-antibiotic era and underscored the importance of developing new classes of antibiotics. The relacidines are a recently discovered group of nonribosomal lipopeptide antibiotics that show promising activity against Gram-negative pathogens and share structural similarities with brevicidine and laterocidine. While the first reports of the relacidines indicated that they possess a C-terminal five-amino acid macrolactone, an N-terminal lipid tail, and an overall positive charge, no stereochemical configuration was assigned, thereby precluding a full structure determination. To address this issue, we here report a bioinformatics guided total synthesis of relacidine A and B and show that the authentic natural products match our predicted and synthesized structures. Following on this, we also synthesized an analogue of relacidine A wherein the ester linkage of the macrolactone was replaced by the corresponding amide. This analogue was found to possess enhanced hydrolytic stability while maintaining the antibacterial activity of the natural product in both in vitro and in vivo efficacy studies.


I. Bioinformatics
We extracted the A-domains from the relacidine, laterocidine and brevicidine BGCs with HMMer (3.3.2) (http://hmmer.org), using the AMP-binding domain profile hidden Markov model (HMM) also used by antiSMASH: https://github.com/antismash/antismash/blob/master/antismash/detection/hmm_detection/data/AMPbinding.hmm We extracted the active sites of these A-domains and predicted their substrates using PARAS (v0.0.1; unpublished; code available at https://pypi.org/project/paras/).We predicted the structures of the Glyrecognizing A-domains of the relacidine, laterocidine and brevicidine BGCs (Table S1) with AlphaFold2, building separate structure models for the N-terminal A-subdomain and the C-terminal Asubdomain.In pymol, we combined the predicted subdomains into complete structures by aligning them to the PDB structure 1AMU, a published X-ray crystallography structure of an A-domain bound to its substrate phenylalanine.We positioned the alanine in the active site of the predicted A-domain structures by mutating the phenylalanine substrate in 1AMU to alanine and measured the distance between the alanine residue and W291/Y290 (or equivalent) with pymol's measurement wizard (Table S1).

Domain
H 2 O : TIPS (95 : 2.5 : 2.5, v/v, 5 mL) for 90 min while shaking.The reaction mixture was filtered through cotton, the filtrate was precipitated from MTBE : petroleum ether (1 : 1, v/v, 45 mL) and centrifuged (4500 rpm, 5 min).The pellet was then resuspended in MTBE : petroleum ether (1 : 1, v/v, 50 mL) and centrifuged again (4500 rpm, 5 min).Finally the pellet containing the crude lipopeptide was dissolved in tBuOH : H 2 O (1 : 1, v/v, 20 mL) and lyophilized overnight.The crude mixtures were subsequently purified by RP-HPLC (See Purification and analysis methods).Fractions were assessed by HPLC and LC-MS and product containing fractions were pooled, frozen and lyophilized to yield the pure lipopeptides as white powders in 3-10% yield over 28 steps.See section V, HPLC and HRMS analysis of peptides, for traces and individual yields.

S7
The reaction mixture was filtered through cotton, the filtrate was precipitated from MTBE : petroleum ether (1 : 1, v/v, 45 mL) and centrifuged (4500 rpm, 5 min).The pellet was then resuspended in MTBE : petroleum ether (1 : 1, v/v, 50 mL) and centrifuged again (4500 rpm, 5 min).Finally the pellet containing the crude lipopeptide was dissolved in tBuOH : H 2 O (1 : 1, v/v, 20 mL) and lyophilized overnight.The crude mixture were subsequently purified by RP-HPLC (See Purification and analysis methods).Fractions were assessed by HPLC and LC-MS and product containing fractions were pooled, frozen and lyophilized to yield the pure lipopeptide as white powder in 18% yield over 30 steps.

V. Purification and analysis methods
Preparative HPLC: Purification was performed on a BESTA-Technik system with a Dr. Maisch ReproSil Gold 120 C18 column (10 µm, 25 x 250 mm) and equipped with a ECOM Flash UV detector.
Runs were performed at a flow rate of 12 mL/min with UV detection at 214 nm and 254 nm.Solvent A = 0.1% TFA in water/MeCN (95 : 5) and solvent B = 0.1% TFA in water/MeCN (5 : 95).A gradient method was employed, starting at 100 % solvent A for 2 min, ramping up to 100 % solvent B over 55 min, remaining at 100 % solvent B for 3 min before ramping down to 100 % solvent A over 1 min and remaining there for 1 min.Product containing fractions were pooled, partially concentrated under vacuum, frozen and then lyophilized to yield pure peptides as white flocculent solids.A small amount of purified peptide was analyzed by analytical HPLC.
Analytical HPLC: Analytical runs were performed on a Shimadzu Prominence-i LC-2030 system with a Dr. Maisch ReproSil Gold 120 C18 ( 5

VI. Culturing conditions and extraction of natural products
Brevibacillus laterosporus MG64 was cultured on Luria-Bertani (LB) agar and colonies grown overnight in 5 ml LB broth at 37°C.This inoculum was transferred to 2 L Erlenmeyer flasks containing 500 ml of LB broth and incubated at 37°C with 220 rpm shaking for 24 h.Cells were collected by centrifugation (10,000 × g, 10 min, 4°C) and extracted with 100 mL of 70% isopropyl alcohol (IPA), pH 2 (acidified with 1 M HCl).Supernatant was separated by centrifugation (6000×g, 10 min, 4 °C) and solvent was evaporated under vacuum using rotary evaporation.The crude extract was reconstituted in H 2 O and filtered with a 0.22 µm syringe filter.were subsequently used to inoculate 5 mL aliquots of TSB that were then incubated at 37 °C.In

Figure S3 .
Figure S2.Hemolytic activity of compounds 3a-5 at 128 µg/mL and 1 h incubation against sheep red blood cells.Colistin was also included as a reference and showed no detectable hemolysis (<0.1%) under the same conditions.

Figure S4 .
Figure S4.HPLC trace showing the reinjection of purified (3a).The peptide eluted as a single peak at 21.588 min using the HPLC method outlined in part.

Figure S5 .
Figure S5.HPLC trace showing the reinjection of purified (3b).The peptide eluted as a single peak at 21.386 min using the HPLC method outlined in part.

Figure S6 .
Figure S6.HPLC trace showing the reinjection of purified (4a).The peptide eluted as a single peak at 22.153 min using the HPLC method outlined in part.

Figure S7 .
Figure S7.HPLC trace showing the reinjection of purified (4b).The peptide eluted as a single peak at 21.990 min using the HPLC method outlined in part.

Figure S8 .Figure S9 .
Figure S8.HPLC trace showing the reinjection of purified (4c).The peptide eluted as a single peak at 22.005 min using the HPLC method outlined in part.

Figure S10 .
Figure S10.HPLC trace showing the reinjection of purified (5).The peptide eluted as a single peak at 21.076 min using the HPLC method outlined in part.

Table S2 .
Minimum inhibitory concentrations (MICs) determined against panel of previously characterized colistinresistant A. baumannii clinical isolates5 a MIC values given in µg/ml.All minimum inhibitory concentrations were determined according to Clinical and StandardsLaboratory Institute (CLSI) guidelines.Blood agar plates were inoculated from glycerol stocks of the different A. baumannii strains used and then incubated for 16 h at 37 °C.Individually grown colonies

Table S3 .
Peptide number, name, chemical formula, exact mass, mass found and overall yield for peptides 3a-5.