Chemically linked phage endolysins broaden and strengthen in vitro bacterial killing

A laboratory study reports that chemically attaching antimicrobial peptides to the phage-derived endolysin LysJEP8 produced protein constructs with stronger bactericidal activity in vitro than the unmodified enzyme. Conjugates made with the peptides GWH1 and PAD improved killing of Acinetobacter baumannii and extended detectable activity to methicillin-resistant Staphylococcus aureus, against which LysJEP8 alone had no detectable growth-inhibitory effect at the tested concentrations. The work, published in Frontiers in Drug Discovery, examines chemical conjugation as an alternative to genetically fusing an endolysin and a peptide into a single recombinant protein.

LysJEP8 originates from Escherichia phage JEP8 and acts by degrading peptidoglycan, the structural layer beneath a bacterial cell wall. Endolysins are being investigated as alternatives to conventional antibiotics because their targets and mechanisms can differ from those of existing drugs, potentially creating opportunities against multidrug-resistant bacteria. Most endolysin research has focused on Gram-positive bacteria, whose peptidoglycan layer is directly exposed to the exterior. Activity against Gram-negative species is more difficult because an outer membrane can prevent an enzyme from reaching its target. LysJEP8 is notable within this field because it showed activity against several Gram-negative ESKAPE pathogens, but it did not inhibit S. aureus in the study’s initial minimum inhibitory concentration tests.

The investigators selected two cationic antimicrobial peptides, GWH1 and PAD, to examine whether pairing them with LysJEP8 could increase potency or broaden the enzyme’s spectrum. Both peptides had previously been characterized as broad-spectrum agents with relatively low haemolytic activity, while PAD had been reported to have strong activity against both Gram-positive and Gram-negative bacteria. The central idea was to make a multivalent endolysin–peptide construct without embedding the highly charged peptide directly within the endolysin’s polypeptide sequence. Such genetic fusions can be difficult to produce because the peptide may interfere with expression, folding, secretion or solubility, resulting in low yields and aggregation.

The researchers instead produced and purified recombinant LysJEP8 in Escherichia coli, synthesized GWH1 and PAD by solid-phase peptide synthesis, and joined the components after production. Sulfo-SMCC served as a heterobifunctional crosslinker: its NHS-ester group reacted with lysine amines on the protein, while its maleimide group coupled to a cysteine added at the C terminus of each peptide. The reaction proceeded in two stages, first activating the protein and then allowing peptide attachment. Unreacted crosslinker and free peptide were removed by desalting and dialysis. This arrangement allowed each component to be manufactured independently before assembly.

Minimum inhibitory concentration assays confirmed that all three agents had strain-dependent activity. LysJEP8 had an MIC of 4 µM against Pseudomonas aeruginosa, 8 µM against A. baumannii and 16 µM against carbapenem-resistant Klebsiella pneumoniae, but no inhibitory activity against S. aureus at concentrations up to 32 µM. GWH1 inhibited S. aureus and A. baumannii at 8 µM, P. aeruginosa at 16 µM and K. pneumoniae at 32 µM. PAD was more consistently active, with MICs of 2 µM against S. aureus, 4 µM against A. baumannii and P. aeruginosa, and 8 µM against K. pneumoniae. These measurements established a baseline for testing whether physical mixtures or covalent conjugates performed differently.

Checkerboard assays showed that simply mixing LysJEP8 with a peptide was not uniformly synergistic. Against S. aureus, neither combination produced synergistic or additive growth inhibition. Against P. aeruginosa, LysJEP8–GWH1 and LysJEP8–PAD were classified as additive. A. baumannii gave the clearest interaction: the LysJEP8–GWH1 combination showed synergy at three tested concentration combinations, whereas LysJEP8–PAD was synergistic at one. In colony-forming-unit assays, the free LysJEP8–GWH1 mixture killed A. baumannii more effectively than either component alone, but the LysJEP8–PAD mixture was no better than its separate components. These results indicated that a non-covalent mixture could sometimes reproduce a combination effect, but it did not consistently provide the strongest killing activity.

Before attaching the peptides to LysJEP8, the team tested the chemistry with bovine serum albumin. SDS-PAGE showed molecular-weight shifts in albumin–GWH1 and albumin–PAD, consistent with covalent attachment. Albumin–PAD caused an approximately one-log reduction in growth of A. baumannii and S. aureus relative to the control and performed comparably to free PAD under the tested conditions, whereas albumin–GWH1 had no detectable antimicrobial effect at those doses. The albumin experiment showed that peptide activity was not necessarily eliminated by conjugation. AlphaFold2 modeling then identified Lys59, Lys75 and Lys129 as the most solvent-exposed lysines in LysJEP8 and therefore the most plausible attachment sites.

Analysis of the LysJEP8 conjugates by SDS-PAGE revealed a mixture of products rather than one uniform molecular species. LysJEP8–GWH1 contained unconjugated protein and forms carrying one, two or three peptides. LysJEP8–PAD had a similar distribution plus a minor population with four peptides. Approximate gel-band analysis indicated that the four-peptide PAD species represented 6.9% of the product, compared with 12.8% unconjugated LysJEP8, although band-intensity percentages are estimates rather than direct measurements of site occupancy. Dynamic light scattering detected an increase in mean hydrodynamic diameter from 3.6 ± 0.1 nm for unmodified LysJEP8 to 5.5 nm for LysJEP8–GWH1 and 5.6 nm for LysJEP8–PAD, independently supporting peptide attachment.

In two-hour colony-forming-unit assays at 37 °C, LysJEP8–PAD showed the greatest bactericidal effect. Against A. baumannii, it reached the assay’s detection limit at 0.5 µM, whereas LysJEP8–GWH1 required 1 µM to reach the same threshold. At 0.5 µM, the bacterial-count difference between unmodified LysJEP8 and LysJEP8–PAD exceeded four orders of magnitude. The free enzyme and free peptides produced only moderate effects at 1 µM by comparison. Against S. aureus, LysJEP8–PAD again reached the detection limit at 0.5 µM, while LysJEP8–GWH1 was less potent but more active than unmodified LysJEP8. The extension of killing to S. aureus was therefore associated primarily with AMP conjugation, which supplied activity the endolysin did not possess on its own.

The results suggest that covalent assembly can produce a different outcome from merely combining soluble components. By linking several peptides to an endolysin, the investigators created multivalent constructs in which membrane-disrupting peptides remain associated with the cell-wall-degrading enzyme. The authors propose that this may help the enzyme reach the bacterial envelope and maintain a higher local concentration of peptide, although that mechanism was not directly verified and requires further testing. The evidence is also limited to laboratory assays: no animal, tissue, pharmacokinetic, toxicity or clinical efficacy study was conducted, and the work does not establish that the conjugates would treat infections in humans. Molecular microheterogeneity, the relationship between particular attachment sites and activity, batch-to-batch consistency and other product characteristics will need evaluation before commercial development.

Chemical conjugation therefore emerged in this study as a practical screening route for combining independently produced endolysins and antimicrobial peptides. It improved the in vitro performance of LysJEP8, particularly when paired with PAD, while avoiding some production problems associated with genetic fusion. The result was not simply a stronger version of the parent enzyme: the conjugates combined the enzyme’s established activity against Gram-negative pathogens with peptide-mediated killing of a Gram-positive pathogen. The work remains a proof-of-concept rather than a finished therapy, but it demonstrates that post-production assembly can provide a flexible way to compare peptide–enzyme combinations and engineer protein-based antimicrobials for further preclinical evaluation.

Subject of Research: In vitro enhancement of phage endolysin bactericidal activity through chemical conjugation with antimicrobial peptides

Article Title: Chemical conjugation of phage endolysins with antimicrobial peptides as a strategy to generate conjugates with enhanced in vitro bactericidal activity

Article References: Carratalá, J. V., Ferrer-Miralles, N., Arís, A., & Garcia-Fruitós, E. (2026). Chemical conjugation of phage endolysins with antimicrobial peptides as a strategy to generate conjugates with enhanced in vitro bactericidal activity. Frontiers in Drug Discovery, 6, Article 1959360. https://doi.org/10.3389/fddsv.2026.1959360

Image Credits: AI Generated

DOI: 10.3389/fddsv.2026.1959360

Keywords: phage endolysins, antimicrobial peptides, chemical conjugation, LysJEP8, bactericidal activity, multidrug-resistant bacteria, Acinetobacter baumannii, Staphylococcus aureus, protein engineering, antimicrobial resistance, Chemical, conjugation