Skip to main content

Advertisement

SpringerLink
Log in

Deep mutational scanning and machine learning for the analysis of antimicrobial-peptide features driving membrane selectivity

  • Article
  • Published:

From Nature Biomedical Engineering

View current issue Submit your manuscript

Abstract

Many antimicrobial peptides directly disrupt bacterial membranes yet can also damage mammalian membranes. It is therefore central to their therapeutic use that rules governing the membrane selectivity of antimicrobial peptides be deciphered. However, this is difficult even for short peptides owing to the large combinatorial space of amino acid sequences. Here we describe a method for measuring the loss or maintenance of antimicrobial-peptide activity for thousands of peptide-sequence variants simultaneously, and its application to Protegrin-1, a potent yet toxic antimicrobial peptide, to determine the positional importance and flexibility of residues across its sequence while identifying variants with changes in membrane selectivity. More bacterially selective variants maintained a membrane-bound secondary structure while avoiding aromatic residues and cysteine pairs. A machine-learning model trained with our datasets accurately predicted membrane-specific activities for over 5.7 million Protegrin-1 variants, and identified one variant that showed substantially reduced toxicity and retention of activity in a mouse model of intraperitoneal infection. The high-throughput methodology may help elucidate sequence–structure–function relationships in antimicrobial peptides and inform the design of peptide-based synthetic drugs.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Log in via an institution

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Institutional subscriptions

Fig. 1: Protegrin-1 deep mutational SLAY predicts residue importance and flexibility.
Fig. 2: Changes in secondary structure correlate with Protegrin-1 lytic activity.
Fig. 3: Membrane selectivity is influenced by aromatics and loss of cysteine pairs.
Fig. 4: Protegrin-1 variants demonstrate strong specificity for bacterial membranes.
Fig. 5: Machine learning identifies mutational profiles promoting membrane specificity.
Fig. 6: Analysis of machine learning performance.

Similar content being viewed by others

Data availability

Raw sequencing data from dmSLAY are available from the SRA database under accession number PRJNA1022479. The raw and analysed datasets generated during the study are available for research purposes from the corresponding author on reasonable request. Source data for the figures are provided with this paper.

Code availability

Coding materials for sequence analysis and machine learning are available on GitHub at https://github.com/facebookresearch/esm#main-models ref. 40 and https://github.com/ziul-bio/DMS_ML_AMP ref. 41.

References

  1. Mookherjee, N., Anderson, M. A., Haagsman, H. P. & Davidson, D. J. Antimicrobial host defence peptides: functions and clinical potential. Nat. Rev. Drug Discov. 19, 311–332 (2020).

    Article  CAS  PubMed  Google Scholar 

  2. Huang, Y., Huang, J. & Chen, Y. Alpha-helical cationic antimicrobial peptides: relationships of structure and function. Protein Cell 1, 143–152 (2010).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  3. Fowler, D. M. & Fields, S. Deep mutational scanning: a new style of protein science. Nat. Methods 11, 801–807 (2014).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  4. Koch, P. et al. Optimization of the antimicrobial peptide Bac7 by deep mutational scanning. BMC Biol. 20, 114 (2022).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  5. Tucker, A. T. et al. Discovery of next-generation antimicrobials through bacterial self-screening of surface-displayed peptide libraries. Cell 172, 618–628.e13 (2018).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  6. Khabbaz, H., Karimi-Jafari, M. H., Saboury, A. A. & BabaAli, B. Prediction of antimicrobial peptides toxicity based on their physico-chemical properties using machine learning techniques. BMC Bioinformatics 22, 549 (2021).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  7. Huang, J. et al. Identification of potent antimicrobial peptides via a machine-learning pipeline that mines the entire space of peptide sequences. Nat. Biomed. Eng. 7, 797–810 (2023).

    Article  CAS  PubMed  Google Scholar 

  8. Randall, J. R. et al. Designing and identifying β-hairpin peptide macrocycles with antibiotic potential. Sci. Adv. 9, eade0008 (2023).

    Article  PubMed  PubMed Central  Google Scholar 

  9. Fahrner, R. L. et al. Solution structure of protegrin-1, a broad-spectrum antimicrobial peptide from porcine leukocytes. Chem. Biol. 3, 543–550 (1996).

    Article  CAS  PubMed  Google Scholar 

  10. Panteleev, P. V., Bolosov, I. A., Balandin, S. V. & Ovchinnikova, T. V. Structure and biological functions of β-hairpin antimicrobial peptides. Acta Nat. 7, 37–47 (2015).

    Article  CAS  Google Scholar 

  11. Steinberg, D. A. et al. Protegrin-1: a broad-spectrum, rapidly microbicidal peptide with in vivo activity. Antimicrob. Agents Chemother. 41, 1738–1742 (1997).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  12. Edwards, I. A. et al. Contribution of amphipathicity and hydrophobicity to the antimicrobial activity and cytotoxicity of β-hairpin peptides. ACS Infect. Dis. 2, 442–450 (2016).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  13. Soundrarajan, N. et al. Protegrin-1 cytotoxicity towards mammalian cells positively correlates with the magnitude of conformational changes of the unfolded form upon cell interaction. Sci. Rep. 9, 11569 (2019).

    Article  PubMed  PubMed Central  Google Scholar 

  14. Díez-Aguilar, M. et al. Murepavadin antimicrobial activity against and resistance development in cystic fibrosis Pseudomonas aeruginosa isolates. J. Antimicrob. Chemother. 76, 984–992 (2021).

    Article  PubMed  Google Scholar 

  15. Moreno-Morales, J., Guardiola, S., Ballesté-Delpierre, C., Giralt, E. & Vila, J. A new synthetic protegrin as a promising peptide with antibacterial activity against MDR Gram-negative pathogens. J. Antimicrob. Chemother. 77, 3077–3085 (2022).

    Article  CAS  PubMed  Google Scholar 

  16. Polyphor Ltd. Pivotal study in nosocomial pneumonia suspected or confirmed to be due to Pseudomonas (PRISM-UDR). Study Record. Beta ClinicalTrials.gov https://beta.clinicaltrials.gov/study/NCT03582007 (2019).

  17. Aumelas, A. et al. Synthesis and solution structure of the antimicrobial peptide protegrin-1. Eur. J. Biochem. 237, 575–583 (1996).

    Article  CAS  PubMed  Google Scholar 

  18. Avitabile, C., D’Andrea, L. D. & Romanelli, A. Circular dichroism studies on the interactions of antimicrobial peptides with bacterial cells. Sci. Rep. 4, 4293 (2014).

    Article  PubMed  PubMed Central  Google Scholar 

  19. Greenfield, N. & Fasman, G. D. Computed circular dichroism spectra for the evaluation of protein conformation. Biochemistry 8, 4108–4116 (1969).

    Article  CAS  PubMed  Google Scholar 

  20. Feng, X. et al. The critical role of tryptophan in the antimicrobial activity and cell toxicity of the duck antimicrobial peptide DCATH. Front. Microbiol. 11, 1146 (2020).

    Article  PubMed  PubMed Central  Google Scholar 

  21. Wei, S. Y. et al. Solution structure of a novel tryptophan-rich peptide with bidirectional antimicrobial activity. J. Bacteriol. 188, 328–334 (2006).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  22. Subbalakshmi, C., Bikshapathy, E., Sitaram, N. & Nagaraj, R. Antibacterial and hemolytic activities of single tryptophan analogs of indolicidin. Biochem. Biophys. Res. Commun. 274, 714–716 (2000).

    Article  CAS  PubMed  Google Scholar 

  23. Azad, M. A. K. et al. Significant accumulation of polymyxin in single renal tubular cells: a medicinal chemistry and triple correlative microscopy approach. Anal. Chem. 87, 1590–1595 (2015).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  24. Sales, G. T. M. & Foresto, R. D. Drug-induced nephrotoxicity. Rev. Assoc. Med. Bras. 66, 82–90 (2020).

    Article  Google Scholar 

  25. Poirel, L., Jayol, A. & Nordmanna, P. Polymyxins: antibacterial activity, susceptibility testing, and resistance mechanisms encoded by plasmids or chromosomes. Clin. Microbiol. Rev. 30, 557–596 (2017).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  26. Bolosov, I. A. et al. Design of protegrin-1 analogs with improved antibacterial selectivity. Pharmaceutics 15, 2047 (2023).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  27. Cherkasov, A. et al. Use of artificial intelligence in the design of small peptide antibiotics effective against a broad spectrum of highly antibiotic-resistant superbugs. ACS Chem. Biol. 4, 65–74 (2009).

    Article  CAS  PubMed  Google Scholar 

  28. Guralp, S. A., Murgha, Y. E., Rouillard, J. M. & Gulari, E. From design to screening: a new antimicrobial peptide discovery pipeline. PLoS ONE 8, e59305 (2013).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  29. Hilpert, K., Winkler, D. F. H. & Hancock, R. E. W. Peptide arrays on cellulose support: SPOT synthesis, a time and cost efficient method for synthesis of large numbers of peptides in a parallel and addressable fashion. Nat. Protoc. 2, 1333–1349 (2007).

    Article  CAS  PubMed  Google Scholar 

  30. Bobone, S. & Stella, L. Selectivity of antimicrobial peptides: a complex interplay of multiple equilibria. Adv. Exp. Med. Biol. 1117, 175–214 (2019).

    Article  CAS  PubMed  Google Scholar 

  31. Lai, J. R., Epand, R. F., Weisblum, B., Epand, R. M. & Gellman, S. H. Roles of salt and conformation in the biological and physicochemical behavior of protegrin-1 and designed analogues: correlation of antimicrobial, hemolytic, and lipid bilayer-perturbing activities. Biochemistry 45, 15718–15730 (2006).

    Article  CAS  PubMed  Google Scholar 

  32. Harwig, S. S. L. Intramolecular disulfide bonds enhance the antimicrobial and lytic activities of protegrins at physiological sodium chloride concentrations. Eur. J. Biochem. 240, 352–357 (1996).

    Article  CAS  PubMed  Google Scholar 

  33. Lai, J. R., Huck, B. R., Weisblum, B. & Gellman, S. H. Design of non-cysteine-containing antimicrobial β-hairpins: structure–activity relationship studies with linear protegrin-1 analogues. Biochemistry 41, 12835–12842 (2002).

    Article  CAS  PubMed  Google Scholar 

  34. Chen, J. et al. Development of protegrins for the treatment and prevention of oral mucositis: structure–activity relationships of synthetic protegrin analogues. Biopolymers 55, 88–98 (2000).

    Article  CAS  PubMed  Google Scholar 

  35. Shen, W., Le, S., Li, Y. & Hu, F. SeqKit: a cross-platform and ultrafast toolkit for FASTA/Q file manipulation. PLoS ONE https://doi.org/10.1371/journal.pone.0163962 (2016).

  36. Dodt, M., Roehr, J. T., Ahmed, R. & Dieterich, C. FLEXBAR—flexible barcode and adapter processing for next-generation sequencing platforms. Biology 1, 895–905 (2012).

    Article  PubMed  PubMed Central  Google Scholar 

  37. Love, M. I., Huber, W. & Anders, S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol. 15, 550 (2014).

    Article  PubMed  PubMed Central  Google Scholar 

  38. Lin, Z. et al. Evolutionary-scale prediction of atomic-level protein structure with a language model. Science https://doi.org/10.1126/science.ade2574 (2023).

  39. Detlefsen, N. S., Hauberg, S. & Boomsma, W. Learning meaningful representations of protein sequences. Nat. Commun. 13, 1914 (2022).

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  40. Facebook Research. Evolutionary scale modeling. GitHub https://github.com/facebookresearch/esm#main-models (2023).

  41. Vieira, L. C. Deep mutational analysis and machine learning uncover antimicrobial peptide features driving membrane selectivity. GitHub https://github.com/ziul-bio/DMS_ML_AMP (2024).

  42. Randall, J. R. et al. Synthetic antibacterial discovery of symbah-1, a macrocyclic β-hairpin peptide antibiotic. iScience 25, 103611 (2021).

    Article  PubMed  PubMed Central  Google Scholar 

Download references

Acknowledgements

We thank the Targeted Therapeutic Drug Discovery and Development Program at the University of Texas for access to circular dichroism training and equipment. The authors disclose support for the research described in this study from the National Institutes of Health (grant numbers AI125337, AI148419 and AI159203), the Welch Foundation (grant number F-2137), the Defense Threat Reduction Agency (grant number ADTRA1-17-C0008) and Tito’s Handmade Vodka.

Author information

Authors and Affiliations

  1. Department of Molecular Biosciences, The University of Texas at Austin, Austin, TX, USA

    Justin R. Randall & Bryan W. Davies

  2. Department of Integrative Biology, The University of Texas at Austin, Austin, TX, USA

    Luiz C. Vieira & Claus O. Wilke

Authors
  1. Justin R. Randall
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Luiz C. Vieira
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Claus O. Wilke
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. Bryan W. Davies
    View author publications

    You can also search for this author in PubMed Google Scholar

Contributions

J.R.R., L.C.V., B.W.D. and C.O.W. conceptualized this work. J.R.R. and L.C.V. were responsible for methodology, investigation and data visualization. B.W.D. and C.O.W. supervised the work and acquired funding. J.R.R. wrote the original draft and L.C.V., B.W.D. and C.O.W. helped with editing.

Corresponding author

Correspondence to Bryan W. Davies.

Ethics declarations

Competing interests

The authors declare no competing interests.

Peer review

Peer review information

Nature Biomedical Engineering thanks Jian Ji and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. Peer reviewer reports are available.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Extended data

Extended Data Fig. 1 Protegin-1 dmSLAY library diversity and sequencing analysis.

a, an alanine scan performed on the native Protegrin-1 (PG-1.0) amino acid sequence showing antibacterial activity (MIC) in µg/ml. Reported MIC is the median of triplicate reactions. b, Chart of the sequence variance found within the Protegrin-1 dmSLAY library. The native Protegrin-1 sequence is shown at the top with mutations observed at each location within the library below. Amino acids are color coded by side chain similarity. Brackets represent where disulfide bonds are present. c, Principal Component Analysis of the overall induced and uninduced triplicate sample variance. d, ROC curve with different MIC cut offs for active and inactivity for log2-fold change cut off across a range of log2-fold change scores (L2FC).

Source data

Extended Data Fig. 2 Selectivity of serine and histidine containing PG-1 variants.

a, Scatter plot of the log2-fold change in MIC versus %Hemolysis for dmSLAY active PG-1 variants from Fig. 3. Dotted line represents PG-1.0 selectivity score. b, Table showing the biochemical characteristics of serine and histidine containing Protegrin-1 variants from dmSLAY. MIC is the median of triplicate assays and %Hemolysis is the mean of triplicate assays. c, Bar chart showing the selectivity score of serine and histidine containing variants on a log2 scale. Residue changes are shown below. Brackets show where disulfide bonds are formed in the native structure.

Source data

Extended Data Fig. 3 Comparing Protegrin-1 variant activity in mixed cultures.

a-d, Graphs of PG-1 (top left), PG-1.1 (bottom left), PG-1.20 (top right), and PG1-37 (bottom right) percentage of bacterial killing (green) and % hemolysis (purple) with 1 × 109 red blood cells (RBC), 1 × 106 E. coli W3110 cells (Bacteria) or both at various concentrations shown on a log2 scale. Each data point is the mean of triplicate reactions and error bars are one standard deviation.

Source data

Extended Data Fig. 4 Training of machine learning models and specific attribute mutational profiles.

a, Precision and recall of for predicting PG-1 variants with an MIC > or < 8 µg/ml. b, predicted versus true hemolysis for trained and test data c, or predicted versus true log10-selectivity score. All models were trained on 80% of data and validated with 20%. Bottom panels: Mutational profiles of variants from 5.7 million candidates with a predicted MIC ≤ 8 µg/ml (a) % hemolysis ≤ 2 (b) or log10-Selectivity score ≤ 0.5 (c).

Source data

Supplementary information

Supplementary Information

Supplementary Methods, Figures and Tables.

Reporting Summary

Peer Review File

Data

Machine-learning-identified Protegin-1 variants.

Source data

Rights and permissions

Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Randall, J.R., Vieira, L.C., Wilke, C.O. et al. Deep mutational scanning and machine learning for the analysis of antimicrobial-peptide features driving membrane selectivity. Nat. Biomed. Eng 8, 842–853 (2024). https://doi.org/10.1038/s41551-024-01243-1

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1038/s41551-024-01243-1

  • Springer Nature Limited

Search

Navigation