Reducing hurdles to clinical trials without compromising the therapeutic promises of peptide candidates becomes an essential step in peptide-based drug design. Machine-learning models are cost-effective and time-saving strategies used to predict biological activities from primary sequences. Their limitations lie in the diversity of peptide sequences and biological information within these models. Additional outlier detection methods are needed to set the boundaries for reliable predictions; the applicability domain. Antimicrobial peptides (AMPs) constitute an extensive library of peptides offering promising avenues against antibiotic-resistant infections. Most AMPs present in clinical trials are administrated topically due to their hemolytic toxicity. Here we developed machine learning models and outlier detection methods that ensure robust predictions for the discovery of AMPs and the design of novel peptides with reduced hemolytic activity. Our best models, gradient boosting classifiers, predicted the hemolytic nature from any peptide sequence with 95–97% accuracy. Nearly 70% of AMPs were predicted as hemolytic peptides. Applying multivariate outlier detection models, we found that 273 AMPs (~ 9%) could not be predicted reliably. Our combined approach led to the discovery of 34 high-confidence non-hemolytic natural AMPs, the de novo design of 507 non-hemolytic peptides, and the guidelines for non-hemolytic peptide design.

Natural products (NPs) are primarily recognized as privileged structures to interact with protein drug targets. Their unique characteristics and structural diversity continue to marvel scientists for developing NP-inspired medicines, even...

No abstract

Chemical analysis of a Didemnum sp. (CMB-01656) collected during scientific Scuba operations off Wasp Island, New South Wales, yielded five new lamellarins A1 (1), A2 (2), A3 (3), A4 (4) and A5 (5) and eight known lamellarins C (6), E (7), K (8), M (9), S (10), T (11), X (12) and χ (13). Analysis of a second Didemnum sp. (CMB-02127) collected during scientific trawling operations along the Northern Rottnest Shelf, Western Australia, yielded the new lamellarin A6 (14) and two known lamellarins G (15) and Z (16). Structures were assigned to 1-16 on the basis of detailed spectroscopic analysis with comparison to literature data and authentic samples. Access to this unique library of natural lamellarins (1-16) provided a rare opportunity for structure-activity relationship (SAR) investigations, probing interactions between lamellarins and the ABC transporter efflux pump P-glycoprotein (P-gp) with a view to reversing multidrug resistance in a human colon cancer cell line (SW620 Ad300). These SAR studies, which were expanded to include the permethylated lamellarin derivative (17) and a series of lamellarin-inspired synthetic coumarins (19-24) and isoquinolines (25-26), successfully revealed 17 as a promising new non-cytotoxic P-gp inhibitor pharmacophore.

Cyclic pentapeptides (e.g. Ac-(cyclo-1,5)-[KAXAD]-NH2 ; X=Ala, 1; Arg, 2) in water adopt one α-helical turn defined by three hydrogen bonds. NMR structure analysis reveals a slight distortion from α-helicity at the C-terminal aspartate caused by torsional restraints imposed by the K(i)-D(i+4) lactam bridge. To investigate this effect on helix nucleation, the more water-soluble 2 was appended to N-, C-, or both termini of a palindromic peptide ARAARAARA (≤5 % helicity), resulting in 67, 92, or 100 % relative α-helicity, as calculated from CD spectra. From the C-terminus of peptides, 2 can nucleate at least six α-helical turns. From the N-terminus, imperfect alignment of the Asp5 backbone amide in 2 reduces helix nucleation, but is corrected by a second unit of 2 separated by 0-9 residues from the first. These cyclic peptides are extremely versatile helix nucleators that can be placed anywhere in 5-25 residue peptides, which correspond to most helix lengths in protein-protein interactions.